From daemon@optimus.ietf.org  Mon Apr  1 06:45:47 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA20056
	for <tsvwg-archive@odin.ietf.org>; Mon, 1 Apr 2002 06:45:41 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA15817
	for tsvwg-archive@odin.ietf.org; Mon, 1 Apr 2002 06:45:44 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA02584;
	Mon, 1 Apr 2002 06:20:36 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA02308
	for <tsvwg@optimus.ietf.org>; Mon, 1 Apr 2002 06:20:23 -0500 (EST)
Received: from stewart.chicago.il.us (host-64-47-24-166.masergy.com [64.47.24.166] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA18067
	for <tsvwg@ietf.org>; Mon, 1 Apr 2002 06:19:59 -0500 (EST)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g31BIso74134;
	Mon, 1 Apr 2002 05:19:04 -0600 (CST)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CA8421D.92AA6334@stewart.chicago.il.us>
Date: Mon, 01 Apr 2002 05:18:54 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: "Iain.Thomson" <Iain.Thomson@uk.thalesgroup.com>
CC: tsvwg@ietf.org, sctpimp@netlab.hcltech.com
Subject: Re: [Tsvwg] Congestion window init
References: <B0B581DB7EDFD2119DC00090273A97F501A5BD0C@nts006.rrl.co.uk>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

"Iain.Thomson" wrote:
> 
> Randall,
> small comment below
> Iain
> =========================
>   Iain Thomson
>   Thales Research Ltd
>   iain.thomson@uk.thalesgroup.com
> =========================
> 
> -----Original Message-----
> From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: Wednesday, March 27, 2002 2:32 PM
> To: Arun Prasad
> Cc: tsvwg@ietf.org; sctpimp@netlab.hcltech.com
> Subject: Re: [Tsvwg] Congestion window init
> 
> Randall wrote:
> >A couple of observations:
> >
> >1) If your peer is limiting you to 1500 bytes no matter if
> >   you are in slow-start or CA you are stuck sending only one
> >   segment at a time.
> >
> >2) You can never grow your cwnd if your peer limits the rwnd
> >   below the value of cwnd.
> 
> With TCP the cwnd grows according to the received ACKs and so may grow very large but the number of bytes that can be sent will still be limited by the size of rwnd. Is SCTP different? I think its RFC 2861 (TCP Congestion Window Validation) that has a useful discussion on this.



Yes, SCTP is a tiny bit  different:

1) You can never grow your cwnd unless you are using all of it.
2) You only are allowed to increase the cwnd by either the
   number of bytes actually ack'd <OR> one P-MTU which ever
   is smaller.

So what will end up happening is that the last ACK that grows the
cwnd may put you slightly over the rwnd (up to P-MTU-1) but after
that it will never grow anymore... since you are never using
the full cwnd....


R
-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Apr  1 16:18:52 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA15327
	for <tsvwg-archive@odin.ietf.org>; Mon, 1 Apr 2002 16:18:51 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA12044
	for tsvwg-archive@odin.ietf.org; Mon, 1 Apr 2002 16:18:54 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA10623;
	Mon, 1 Apr 2002 15:51:11 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA10597
	for <tsvwg@optimus.ietf.org>; Mon, 1 Apr 2002 15:51:09 -0500 (EST)
Received: from mg03.austin.ibm.com (mg03.austin.ibm.com [192.35.232.20])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA14218
	for <tsvwg@ietf.org>; Mon, 1 Apr 2002 15:51:06 -0500 (EST)
Received: from austin.ibm.com (netmail1.austin.ibm.com [9.3.7.138])
	by mg03.austin.ibm.com (AIX4.3/8.9.3/8.9.3) with ESMTP id OAA08756
	for <tsvwg@ietf.org>; Mon, 1 Apr 2002 14:47:14 -0600
Received: from austin.ibm.com (ambika.austin.ibm.com [9.53.150.77])
	by austin.ibm.com (AIX4.3/8.9.3/8.9.3) with ESMTP id OAA05616
	for <tsvwg@ietf.org>; Mon, 1 Apr 2002 14:51:07 -0600
Message-ID: <3CA8C83B.52714929@austin.ibm.com>
Date: Mon, 01 Apr 2002 14:51:07 -0600
From: venkat venkatsubra <venkats@austin.ibm.com>
Organization: IBM
X-Mailer: Mozilla 4.76iC-CCK-MCD  [en_US] (X11; U; AIX 5.1)
X-Accept-Language: en
MIME-Version: 1.0
To: tsvwg@ietf.org
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] delayed SACK when sender is retransmitting
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Dear All,

When a receiver receives a retransmitted DATA chunk that advances
the cumulative TSN Ack number, isn't it good to send the SACK
immediately
instead of delaying it ? Otherwise what we see is the sender when
retransmitting stalling waiting for this SACK before it can proceed
further.
Since the rfc (2960) requires you to SACK immediately on receiving
the first DATA chunk, a similar rule could help in case of receiving
retransmitted DATA chunks i.e. on receiving a DATA chunk that fills a
hole and advances the cumulative TSN Ack.  Or,  is this left to the
implementation ?
Thanks!

Venkat


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Apr  1 19:09:09 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA19790
	for <tsvwg-archive@odin.ietf.org>; Mon, 1 Apr 2002 19:09:09 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id TAA21640
	for tsvwg-archive@odin.ietf.org; Mon, 1 Apr 2002 19:09:11 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA20367;
	Mon, 1 Apr 2002 18:43:44 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA20336
	for <tsvwg@optimus.ietf.org>; Mon, 1 Apr 2002 18:43:42 -0500 (EST)
Received: from stewart.chicago.il.us (host-64-47-24-166.masergy.com [64.47.24.166] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA19119
	for <tsvwg@ietf.org>; Mon, 1 Apr 2002 18:43:39 -0500 (EST)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g31Nhfo75499
	for <tsvwg@ietf.org>; Mon, 1 Apr 2002 17:43:42 -0600 (CST)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CA8F0AD.1F864E52@stewart.chicago.il.us>
Date: Mon, 01 Apr 2002 17:43:41 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: TSV WG list <tsvwg@ietf.org>
Subject: Re: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt]
References: <3CA8EFB3.BC083B30@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hmm..

Since this did not go to the working group in my reply-all
and my second forward also failed (I think due to anti-spamming)
I will retry.... forwarding to my home email first :>

More comments on the wonderful sctpcsum issue :>

R


Randall Stewart at work wrote:
> 
> I have just looked over the whole thread... and I see both points
> of view on this matter... I must agree Jonathan's arguments are (as
> usual)
> quite persuasive.
> 
> Now another thought I had ...  Where did this wording come from?
> 
> So I went back to RFC2960 and what do I find?
> 
> "
>    When an SCTP packet is received, the receiver MUST first check the
>    Adler-32 checksum:
> 
>    1) Store the received Adler-32 checksum value aside,
> 
> Stewart, et al.             Standards Track                    [Page 82]
> 
> RFC 2960          Stream Control Transmission Protocol      October 2000
> 
>    2) Replace the 32 bits of the checksum field in the received SCTP
>       packet with all '0's and calculate an Adler-32 checksum value of
>       the whole received packet.  And,
> 
>    3) Verify that the calculated Adler-32 checksum is the same as the
>       received Adler-32 checksum.  If not, the receiver MUST treat the
>       packet as an invalid SCTP packet.
> 
>    The default procedure for handling invalid SCTP packets is to
>    silently discard them.
> "
> 
> Looks kind of familiar does it not?
> 
> Now I think (if I have read correctly) what Jonathan has proposed is
> something along
> the lines of:
> 
> "
>     After the packet is constructed (containing the SCTP common header and
>     one or more control or DATA chunks), the transmitter MUST do the
>     following:
> 
>     1) Fill in the proper Verification Tag in the SCTP common header and
>        initialize the Checksum field to 0's.
> 
>     2) Calculate the CRC-32c of the whole packet, including the SCTP common
>        header and all the chunks.
> 
>     3) Put the resultant value into the Checksum field in the common header,
>        and leave the rest of the bits unchanged.
> 
>     When an SCTP packet is received, the receiver MUST perform the checksum
>     algorithm. It SHOULD perform this algorithm using the following procedure:
> 
>     1) Store the received CRC-32c value,
> 
>     2) Replace the 32 bits of the Checksum field in the received SCTP packet
>        with all '0's and calculate a CRC-32c value of the whole received
>        packet.  And,
> 
>     3) Verify that the calculated CRC-32c value is the same as the received
>        CRC-32c value.  If not, the receiver MUST treat the packet as an
>        invalid SCTP packet.
> 
>     A fully conformant SCTP implementation SHOULD use this exact procedure.
> "
> 
> Jonathan: does this capture your AS IF <current> spec sentiment?
> 
> Or did I miss it?
> 
> Would this be acceptable to all concerned? Or am I off base here?
> 
> R
> --
> Randall R. Stewart
> ITD
> Cisco Systems Inc.
> rrs@cisco.com 815-342-5222(cell) or 815-477-2127(office)


Now... I think there are two ways to go.. if this captured
Jonathans idea and the working group agrees... we insert
the above... Or just leave the text as is... I am fine
with either method :>

I have -05 ready with leaving things the way they are...

Comments most welcome :-)

R

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Apr  1 22:26:03 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA23938
	for <tsvwg-archive@odin.ietf.org>; Mon, 1 Apr 2002 22:26:02 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id WAA00816
	for tsvwg-archive@odin.ietf.org; Mon, 1 Apr 2002 22:26:04 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id VAA28602;
	Mon, 1 Apr 2002 21:41:02 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id VAA28574
	for <tsvwg@optimus.ietf.org>; Mon, 1 Apr 2002 21:41:01 -0500 (EST)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA23179
	for <tsvwg@ietf.org>; Mon, 1 Apr 2002 21:40:55 -0500 (EST)
Received: from netlab.hcltech.com (arun.hclt-guindy.co.in [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id IAA20004;
	Tue, 2 Apr 2002 08:31:33 +0530
Message-ID: <3CA91A57.DB682B2@netlab.hcltech.com>
Date: Tue, 02 Apr 2002 08:11:27 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: venkat venkatsubra <venkats@austin.ibm.com>
CC: tsvwg@ietf.org
Subject: Re: [Tsvwg] delayed SACK when sender is retransmitting
References: <3CA8C83B.52714929@austin.ibm.com>
Content-Type: multipart/alternative;
 boundary="------------7EBE43325211614336608A92"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


--------------7EBE43325211614336608A92
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

hi Venkat,
    comments inline....

venkat venkatsubra wrote:

> Dear All,
>
> When a receiver receives a retransmitted DATA chunk that advances
> the cumulative TSN Ack number, isn't it good to send the SACK
> immediately
> instead of delaying it ? Otherwise what we see is the sender when
> retransmitting stalling waiting for this SACK before it can proceed
> further.

        How will the receiver know whether a Data chunk is retransmitted chunk???
We can say a data chunk received might be a retransmitted one, for the following
conditions.... But in all the below conditions a SACK is to be generated according
to Rfc....


1) GAP:
"6.7 Report Gaps in Received DATA TSNs

   Upon the reception of a new DATA chunk, an endpoint shall examine the
   continuity of the TSNs received.  If the endpoint detects a gap in
   the received DATA chunk sequence, it SHOULD send a SACK with Gap Ack
   Blocks immediately.   "

2)DUPLICATE:
"   When a packet arrives with duplicate DATA chunk(s) and with no new
   DATA chunk(s), the endpoint MUST immediately send a SACK with no
   delay.  If a packet arrives with duplicate DATA chunk(s) bundled with
   new DATA chunks, the endpoint MAY immediately send a SACK. "




>
> Since the rfc (2960) requires you to SACK immediately on receiving
> the first DATA chunk, a similar rule could help in case of receiving
> retransmitted DATA chunks i.e. on receiving a DATA chunk that fills a
> hole and advances the cumulative TSN Ack.  Or,  is this left to the
> implementation ?
> Thanks!
>
> Venkat
>

Thanks
-arun

>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************


--------------7EBE43325211614336608A92
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
hi Venkat,
<br>&nbsp;&nbsp;&nbsp; comments inline....
<p>venkat venkatsubra wrote:
<blockquote TYPE=CITE>Dear All,
<p>When a receiver receives a retransmitted DATA chunk that advances
<br>the cumulative TSN Ack number, isn't it good to send the SACK
<br>immediately
<br>instead of delaying it ? Otherwise what we see is the sender when
<br>retransmitting stalling waiting for this SACK before it can proceed
<br>further.</blockquote>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; How will the receiver know whether
a Data chunk is retransmitted chunk???
<br>We can say a data chunk received might be a retransmitted one, for
the following
<br>conditions.... But in all the below conditions a SACK is to be generated
according
<br>to Rfc....
<br>&nbsp;
<p><b>1) GAP:</b>
<br><b>"6.7 Report Gaps in Received DATA TSNs</b><b></b>
<p><b>&nbsp;&nbsp; Upon the reception of a new DATA chunk, an endpoint
shall examine the</b>
<br><b>&nbsp;&nbsp; continuity of the TSNs received.&nbsp; If the endpoint
detects a gap in</b>
<br><b>&nbsp;&nbsp; the received DATA chunk sequence, it SHOULD send a
SACK with Gap Ack</b>
<br><b>&nbsp;&nbsp; Blocks immediately.&nbsp;&nbsp; "</b><b></b>
<p><b>2)DUPLICATE:</b>
<br><b>"&nbsp;&nbsp; When a packet arrives with duplicate DATA chunk(s)
and with no new</b>
<br><b>&nbsp;&nbsp; DATA chunk(s), the endpoint MUST immediately send a
SACK with no</b>
<br><b>&nbsp;&nbsp; delay.&nbsp; If a packet arrives with duplicate DATA
chunk(s) bundled with</b>
<br><b>&nbsp;&nbsp; new DATA chunks, the endpoint MAY immediately send
a SACK. "</b>
<br>&nbsp;
<br>&nbsp;
<br><b></b>&nbsp;
<blockquote TYPE=CITE>&nbsp;
<br>Since the rfc (2960) requires you to SACK immediately on receiving
<br>the first DATA chunk, a similar rule could help in case of receiving
<br>retransmitted DATA chunks i.e. on receiving a DATA chunk that fills
a
<br>hole and advances the cumulative TSN Ack.&nbsp; Or,&nbsp; is this left
to the
<br>implementation ?
<br>Thanks!
<p>Venkat
<br>&nbsp;</blockquote>

<p><br>Thanks
<br>-arun
<blockquote TYPE=CITE>&nbsp;
<br>_______________________________________________
<br>tsvwg mailing list
<br>tsvwg@ietf.org
<br><a href="https://www1.ietf.org/mailman/listinfo/tsvwg">https://www1.ietf.org/mailman/listinfo/tsvwg</a></blockquote>

<p>--
<br>****************************************************************
<br>V.Arun Prasad
<br>HCL Technologies Ltd.
<br>51, Jawaharlal Nehru Road,
<br>Ekkattuthangal,
<br>Guindy Industrial Estate,
<br>Chennai - 600097.
<p>Contact # : 9144 - 2334174
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
9144 - 2334181
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
extn : 233
<br>****************************************************************
<br>&nbsp;</html>

--------------7EBE43325211614336608A92--


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr  2 07:37:17 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA13247
	for <tsvwg-archive@odin.ietf.org>; Tue, 2 Apr 2002 07:37:17 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA08622
	for tsvwg-archive@odin.ietf.org; Tue, 2 Apr 2002 07:37:18 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA06983;
	Tue, 2 Apr 2002 07:16:57 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA06952
	for <tsvwg@optimus.ietf.org>; Tue, 2 Apr 2002 07:16:55 -0500 (EST)
Received: from stewart.chicago.il.us (host-64-47-24-166.masergy.com [64.47.24.166] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA12474
	for <tsvwg@ietf.org>; Tue, 2 Apr 2002 07:16:52 -0500 (EST)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g32CGio76769;
	Tue, 2 Apr 2002 06:16:44 -0600 (CST)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CA9A12C.BE380CE2@stewart.chicago.il.us>
Date: Tue, 02 Apr 2002 06:16:44 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Arun Prasad <arun@netlab.hcltech.com>
CC: venkat venkatsubra <venkats@austin.ibm.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] delayed SACK when sender is retransmitting
References: <3CA8C83B.52714929@austin.ibm.com> <3CA91A57.DB682B2@netlab.hcltech.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Venkat:

I think Arun has mad the correct observation here.

It is really impossible to tell the difference between
a retransmission and a network delayed packet. The
receiver has no idea (even when a GAP exists) if 
the packets are just received out of order OR the
GAP filling packet is a fast-retransmit or T3-Timeout.

Adding the 200ms delay I don't think will be a problem since:

1) If it is a retransmission then there is a good chance
   the network is related to congestion and slowing down
   might not be such a bad thing.

2) If it is a normal but re-ordered transmission the peer
   may be doing a RTO measurement on the packet and getting
   the 200ms delay put into the time recorded (the normal
   delay) is a good thing.

All in all it is a good thing :>

R

Arun Prasad wrote:
> 
> hi Venkat,
>     comments inline....
> 
> venkat venkatsubra wrote:
> 
> > Dear All,
> >
> > When a receiver receives a retransmitted DATA chunk that advances
> > the cumulative TSN Ack number, isn't it good to send the SACK
> > immediately
> > instead of delaying it ? Otherwise what we see is the sender when
> > retransmitting stalling waiting for this SACK before it can proceed
> > further.
> 
>         How will the receiver know whether a Data chunk is
> retransmitted chunk???
> We can say a data chunk received might be a retransmitted one, for the
> following
> conditions.... But in all the below conditions a SACK is to be
> generated according
> to Rfc....
> 
> 
> 1) GAP:
> "6.7 Report Gaps in Received DATA TSNs
> 
>    Upon the reception of a new DATA chunk, an endpoint shall examine
> the
>    continuity of the TSNs received.  If the endpoint detects a gap in
>    the received DATA chunk sequence, it SHOULD send a SACK with Gap
> Ack
>    Blocks immediately.   "
> 
> 2)DUPLICATE:
> "   When a packet arrives with duplicate DATA chunk(s) and with no new
> 
>    DATA chunk(s), the endpoint MUST immediately send a SACK with no
>    delay.  If a packet arrives with duplicate DATA chunk(s) bundled
> with
>    new DATA chunks, the endpoint MAY immediately send a SACK. "
> 
> 
> 
> 
> >
> > Since the rfc (2960) requires you to SACK immediately on receiving
> > the first DATA chunk, a similar rule could help in case of receiving
> >
> > retransmitted DATA chunks i.e. on receiving a DATA chunk that fills
> > a
> > hole and advances the cumulative TSN Ack.  Or,  is this left to the
> > implementation ?
> > Thanks!
> >
> > Venkat
> >
> 
> Thanks
> -arun
> 
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> --
> ****************************************************************
> V.Arun Prasad
> HCL Technologies Ltd.
> 51, Jawaharlal Nehru Road,
> Ekkattuthangal,
> Guindy Industrial Estate,
> Chennai - 600097.
> 
> Contact # : 9144 - 2334174
>             9144 - 2334181
>             extn : 233
> ****************************************************************
> 

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr  2 11:08:13 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA20703
	for <tsvwg-archive@odin.ietf.org>; Tue, 2 Apr 2002 11:08:13 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA21432
	for tsvwg-archive@odin.ietf.org; Tue, 2 Apr 2002 11:08:14 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA19921;
	Tue, 2 Apr 2002 10:48:38 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA19891
	for <tsvwg@optimus.ietf.org>; Tue, 2 Apr 2002 10:48:36 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA19727
	for <tsvwg@ietf.org>; Tue, 2 Apr 2002 10:48:33 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir04nok.ntc.nokia.com (esvir04nokt.ntc.nokia.com [172.21.143.36])
	by mgw-x2.nokia.com (Switch-2.2.1/Switch-2.2.0) with ESMTP id g32Fmo516240
	for <tsvwg@ietf.org>; Tue, 2 Apr 2002 18:48:51 +0300 (EET DST)
Received: from esebh003.NOE.Nokia.com (unverified) by esvir04nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5a046d1805ac158f24077@esvir04nok.ntc.nokia.com>;
 Tue, 2 Apr 2002 18:48:35 +0300
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh003.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Tue, 2 Apr 2002 18:48:34 +0300
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] delayed SACK when sender is retransmitting
Date: Tue, 2 Apr 2002 18:48:34 +0300
Message-ID: <58AEDE27D48F884285007ABF827FF26308EF46@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] delayed SACK when sender is retransmitting
Thread-Index: AcHZvv1vrZbiqlGLTKyLehi6N4o1LQAnepEw
To: <venkats@austin.ibm.com>, <tsvwg@ietf.org>
X-OriginalArrivalTime: 02 Apr 2002 15:48:34.0957 (UTC) FILETIME=[D8F0FBD0:01C1DA5D]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id KAA19892
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi Venkat!

> -----Original Message-----
> From: ext venkat venkatsubra [mailto:venkats@austin.ibm.com]
> Sent: 01. April 2002 23:51
> To: tsvwg@ietf.org
> Subject: [Tsvwg] delayed SACK when sender is retransmitting
> 
> 
> Dear All,
> 
> When a receiver receives a retransmitted DATA chunk that advances
> the cumulative TSN Ack number, isn't it good to send the SACK
> immediately
> instead of delaying it ? Otherwise what we see is the sender when
> retransmitting stalling waiting for this SACK before it can proceed
> further.

	But, doing so, you would never delay SACKs at all... When you receive a DATA chunk, either it advances the Cumulative TSN, creates a hole, or it is duplicated... In the last two options, RFC 2960 already states that you cannot delay the SACK...

	And now that I re-read your paragraph above, how can the receiver know that the DATA has been retransmitted? It will know it only if it is duplicated, in which case it will send the SACK (providing no new data is bundled with the duplicated chunk)...

> Since the rfc (2960) requires you to SACK immediately on receiving
> the first DATA chunk, a similar rule could help in case of receiving
> retransmitted DATA chunks i.e. on receiving a DATA chunk that fills a
> hole and advances the cumulative TSN Ack.  Or,  is this left to the
> implementation ?

	This could have also happened due to reordering of packets in the network...

	I think, that in case of a retransmission (due to T3 expiration) the data sender will only send one DATA chunk, a duplicated one, and thus the data receiver will immediately send the SACK you want to receive...

	I hope this helps...

	BR Iván Arias Rodríguez

> Thanks!
> 
> Venkat
> 
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr  2 11:17:53 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA21214
	for <tsvwg-archive@odin.ietf.org>; Tue, 2 Apr 2002 11:17:53 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA22068
	for tsvwg-archive@odin.ietf.org; Tue, 2 Apr 2002 11:17:54 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA21079;
	Tue, 2 Apr 2002 11:02:19 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA21049
	for <tsvwg@optimus.ietf.org>; Tue, 2 Apr 2002 11:02:16 -0500 (EST)
Received: from mg03.austin.ibm.com (mg03.austin.ibm.com [192.35.232.20])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA20366
	for <tsvwg@ietf.org>; Tue, 2 Apr 2002 11:02:14 -0500 (EST)
Received: from austin.ibm.com (netmail1.austin.ibm.com [9.3.7.138])
	by mg03.austin.ibm.com (AIX4.3/8.9.3/8.9.3) with ESMTP id JAA32108;
	Tue, 2 Apr 2002 09:58:20 -0600
Received: from austin.ibm.com (ambika.austin.ibm.com [9.53.150.77])
	by austin.ibm.com (AIX4.3/8.9.3/8.9.3) with ESMTP id KAA32440;
	Tue, 2 Apr 2002 10:02:14 -0600
Message-ID: <3CA9D606.7E0C3810@austin.ibm.com>
Date: Tue, 02 Apr 2002 10:02:14 -0600
From: venkat venkatsubra <venkats@austin.ibm.com>
Organization: IBM
X-Mailer: Mozilla 4.76iC-CCK-MCD  [en_US] (X11; U; AIX 5.1)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: tsvwg@ietf.org
Subject: Re: [Tsvwg] delayed SACK when sender is retransmitting
References: <58AEDE27D48F884285007ABF827FF26308EF46@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Ivan,

Ivan.Arias-Rodriguez@nokia.com wrote:

>         And now that I re-read your paragraph above, how can the receiver know that the DATA has been retransmitted? It will know it only if it is duplicated, in which case it will send the SACK (providing no new data is bundled with the duplicated chunk)...

Receiver need know. For the receiver it is either a retransmitted
or reordered DATA chunk.  But Randall feels delaying the SACK
in both cases is good.

>         I hope this helps...
>
>         BR Iván Arias Rodríguez
>
> > Thanks!

Venkat


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr  2 11:25:53 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA21537
	for <tsvwg-archive@odin.ietf.org>; Tue, 2 Apr 2002 11:25:53 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA22402
	for tsvwg-archive@odin.ietf.org; Tue, 2 Apr 2002 11:25:55 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA21750;
	Tue, 2 Apr 2002 11:14:00 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA21717
	for <tsvwg@optimus.ietf.org>; Tue, 2 Apr 2002 11:13:56 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA20938
	for <tsvwg@ietf.org>; Tue, 2 Apr 2002 11:13:54 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir04nok.ntc.nokia.com (esvir04nokt.ntc.nokia.com [172.21.143.36])
	by mgw-x2.nokia.com (Switch-2.2.1/Switch-2.2.0) with ESMTP id g32GE9527929
	for <tsvwg@ietf.org>; Tue, 2 Apr 2002 19:14:09 +0300 (EET DST)
Received: from esebh003.NOE.Nokia.com (unverified) by esvir04nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5a0484437bac158f24077@esvir04nok.ntc.nokia.com>;
 Tue, 2 Apr 2002 19:13:53 +0300
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh003.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Tue, 2 Apr 2002 19:13:53 +0300
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] delayed SACK when sender is retransmitting
Date: Tue, 2 Apr 2002 19:13:53 +0300
Message-ID: <58AEDE27D48F884285007ABF827FF26308EF47@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] delayed SACK when sender is retransmitting
Thread-Index: AcHaX8c6UnhVh9KaTLKEdYTV1wMAagAACDMQ
To: <venkats@austin.ibm.com>
Cc: <tsvwg@ietf.org>
X-OriginalArrivalTime: 02 Apr 2002 16:13:53.0358 (UTC) FILETIME=[61FA96E0:01C1DA61]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id LAA21718
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi again!

> Ivan,
> 
> Ivan.Arias-Rodriguez@nokia.com wrote:
> 
> >         And now that I re-read your paragraph above, how 
> can the receiver know that the DATA has been retransmitted? 
> It will know it only if it is duplicated, in which case it 
> will send the SACK (providing no new data is bundled with the 
> duplicated chunk)...
> 
> Receiver need know. For the receiver it is either a retransmitted
> or reordered DATA chunk.  But Randall feels delaying the SACK
> in both cases is good.

	But there is no way the receiver can know if it is the original DATA chunk or if it has been resent because it was lost in the network... The fact of receiving a TSN that fills a hole might be caused either by packet loss and retransmission, or by reordering... In the second case you have already been sending the SACKs for every DATA chunk previously received and you have not lost the ACK clock... so you might receive more DATA chunks that will trigger the sending of the SACK...

	If the packet was lost, then I don't think that an extra 200ms delay is really that important (since you will drastically lower your sending capabilities in this case). Besides, if the packet was not lost (maybe the SACK was lost, not the DATA chunk), and the data sender is retransmitting, you would normally include just duplicate TSNs in that first SCTP datagram, and thus, the receiver will ack it immediately...

	In any case, I don't think it is a big problem... And in the normal case, SACK delay saves both bandwidth and processing time at the receiver/routers...

	Bye!

	BR Iván Arias Rodríguez

> >         I hope this helps...
> >
> >         BR Iván Arias Rodríguez
> >
> > > Thanks!
> 
> Venkat
> 
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Apr  2 14:47:20 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA29827
	for <tsvwg-archive@odin.ietf.org>; Tue, 2 Apr 2002 14:47:19 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA08586
	for tsvwg-archive@odin.ietf.org; Tue, 2 Apr 2002 14:47:14 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA07436;
	Tue, 2 Apr 2002 14:28:05 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA07409
	for <tsvwg@optimus.ietf.org>; Tue, 2 Apr 2002 14:28:02 -0500 (EST)
Received: from stewart.chicago.il.us (host-64-47-24-166.masergy.com [64.47.24.166] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA29061
	for <tsvwg@ietf.org>; Tue, 2 Apr 2002 14:28:00 -0500 (EST)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g32J7So77495;
	Tue, 2 Apr 2002 13:07:28 -0600 (CST)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CAA016F.15FCB558@stewart.chicago.il.us>
Date: Tue, 02 Apr 2002 13:07:28 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: mallman@grc.nasa.gov
CC: Arun Prasad <arun@netlab.hcltech.com>,
        venkat venkatsubra <venkats@austin.ibm.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] delayed SACK when sender is retransmitting
References: <200204021605.LAA21857@guns.lerc.nasa.gov>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Mark Allman wrote:
> 
> > I think Arun has mad the correct observation here.
> >
> > It is really impossible to tell the difference between
> > a retransmission and a network delayed packet.
> 
> Yeah, I agree.
> 
> But, I think the packet should be immediatly SACKed.  In RFC 2581 we
> added some language that says you should immedaitly ACK packets that
> either create a hole in the sequence (or TSN, in this case) space or
> that fill in the hole (or part of a hole).  Doing so provides the
> quickest possible feedback to the sender as to the state of the
> world -- which is a good thing for loss recovery algorithms.
> 
Hmm...

So what you are proposing is to always send a SACK without
delay when there are other TSN's ahead of the one that 
advances the cum-ack point. This seems ok.. I don't see that
it is prohibited explicitly or on the other hand encouraged to
do so... in fact I doubt that I did it in our bsd implementation
I will have to add this :>

R


-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Apr  2 14:49:09 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA29910
	for <tsvwg-archive@odin.ietf.org>; Tue, 2 Apr 2002 14:49:09 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA08727
	for tsvwg-archive@odin.ietf.org; Tue, 2 Apr 2002 14:49:10 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA07392;
	Tue, 2 Apr 2002 14:27:38 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA07365
	for <tsvwg@optimus.ietf.org>; Tue, 2 Apr 2002 14:27:36 -0500 (EST)
Received: from stewart.chicago.il.us (host-64-47-24-166.masergy.com [64.47.24.166] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA29057
	for <tsvwg@ietf.org>; Tue, 2 Apr 2002 14:27:34 -0500 (EST)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g32JC6o77514;
	Tue, 2 Apr 2002 13:12:06 -0600 (CST)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CAA0285.59A49DD2@stewart.chicago.il.us>
Date: Tue, 02 Apr 2002 13:12:06 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: venkat venkatsubra <venkats@austin.ibm.com>
CC: Ivan.Arias-Rodriguez@nokia.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] delayed SACK when sender is retransmitting
References: <58AEDE27D48F884285007ABF827FF26308EF46@esebe017.NOE.Nokia.com> <3CA9D606.7E0C3810@austin.ibm.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

venkat venkatsubra wrote:

> Receiver need know. For the receiver it is either a retransmitted
> or reordered DATA chunk.  But Randall feels delaying the SACK
> in both cases is good.
> 


The trick is that you the receiver can NOT know it was
a retransmit. Mark thinks its ok to SACK right away IF and
only IF you have a gap that is being filled in.

So, for example, one could see the following:

Init-TSN= X

  <------TSN=X+1
  ---->SACK (gap no X)
  <------TSN=X+2
  ---->SACK (gap no X)
  <------TSN=X
  ---->SACK (X fills a hole since X+1 and X+2 are here)-->

Now the trick to this is being able to detect that YES
I have TSN's greater than X when X arrives.

I don't see that if you delay things are hurt either.


R


  
-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Apr  2 15:13:20 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA01033
	for <tsvwg-archive@odin.ietf.org>; Tue, 2 Apr 2002 15:13:19 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA10546
	for tsvwg-archive@odin.ietf.org; Tue, 2 Apr 2002 15:13:21 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA09145;
	Tue, 2 Apr 2002 14:54:39 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA09096
	for <tsvwg@ns.ietf.org>; Tue, 2 Apr 2002 14:54:34 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA00102
	for <tsvwg@ietf.org>; Tue, 2 Apr 2002 14:54:32 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir04nok.ntc.nokia.com (esvir04nokt.ntc.nokia.com [172.21.143.36])
	by mgw-x2.nokia.com (Switch-2.2.1/Switch-2.2.0) with ESMTP id g32Jsm522102
	for <tsvwg@ietf.org>; Tue, 2 Apr 2002 22:54:48 +0300 (EET DST)
Received: from esebh003.NOE.Nokia.com (unverified) by esvir04nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5a054e4656ac158f24077@esvir04nok.ntc.nokia.com>;
 Tue, 2 Apr 2002 22:54:32 +0300
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh003.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Tue, 2 Apr 2002 22:54:32 +0300
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] delayed SACK when sender is retransmitting 
Date: Tue, 2 Apr 2002 22:54:32 +0300
Message-ID: <58AEDE27D48F884285007ABF827FF26308EF4A@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] delayed SACK when sender is retransmitting 
Thread-Index: AcHafxAU0Ewxb4B1QLyWDhVHD8MyggAABQxQ
To: <mallman@grc.nasa.gov>, <randall@stewart.chicago.il.us>
Cc: <venkats@austin.ibm.com>, <tsvwg@ietf.org>
X-OriginalArrivalTime: 02 Apr 2002 19:54:32.0358 (UTC) FILETIME=[3509AC60:01C1DA80]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id OAA09101
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

> > Mark thinks its ok to SACK right away IF and only IF you have a
> > gap that is being filled in.
> 
> Almost.
> 
> Mark thinks its OK to SACK immediatly if we are filling in all or
> part of a gap **or** the incoming segment creates or enlarges a gap.

	I guess that you are mostly saying the same, as in the second and third cases (filling part of a gap, or creating a new one), you will end up having gaps, and thus you will send the SACK anyway...

	So the only new case would be that if you had gaps before the receipt of the datagram, no matter if after processing the datagram you don't have them, you must always acknowledge the receipt of those chunks immediately...

	So I think that the basic idea here is that, apart from the case in which you receive only duplicate DATA chunks, you override the delayed SACK rule if before or after processing the DATA chunks you have gaps in your incoming TSN sequence... right?

	BR Iván Arias Rodríguez

> From RFC 2581:
> 
>    Out-of-order data segments SHOULD be acknowledged immediately, in
>    order to accelerate loss recovery.  To trigger the fast retransmit
>    algorithm, the receiver SHOULD send an immediate duplicate ACK when
>    it receives a data segment above a gap in the sequence space.  To
>    provide feedback to senders recovering from losses, the receiver
>    SHOULD send an immediate ACK when it receives a data segment that
>    fills in all or part of a gap in the sequence space.
> 
> allman
> 
> 
> --
> Mark Allman -- BBN/NASA GRC -- http://roland.grc.nasa.gov/~mallman/
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Wed Apr  3 02:11:10 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA26231
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 02:11:10 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id CAA29379
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 02:11:13 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA27786;
	Wed, 3 Apr 2002 01:45:38 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA27757
	for <tsvwg@ns.ietf.org>; Wed, 3 Apr 2002 01:45:36 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@[209.44.42.193])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA17659
	for <tsvwg@ietf.org>; Wed, 3 Apr 2002 01:45:31 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id AAA19061;
	Wed, 3 Apr 2002 00:45:21 -0600
Date: Wed, 3 Apr 2002 00:45:20 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: venkat venkatsubra <venkats@austin.ibm.com>
Cc: tsvwg@ietf.org
Subject: Re: [Tsvwg] delayed SACK when sender is retransmitting
Message-ID: <20020403004520.A18869@openss7.org>
Reply-To: bidulock@openss7.org
References: <3CA8C83B.52714929@austin.ibm.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <3CA8C83B.52714929@austin.ibm.com>; from venkats@austin.ibm.com on Mon, Apr 01, 2002 at 02:51:07PM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Venkat,

On Mon, 01 Apr 2002, venkat venkatsubra wrote:

> Dear All,
> 
> When a receiver receives a retransmitted DATA chunk that advances
> the cumulative TSN Ack number, isn't it good to send the SACK
> immediately
> instead of delaying it ? Otherwise what we see is the sender when
> retransmitting stalling waiting for this SACK before it can proceed
> further.

Why is your retransmitting sender stalling waiting to
receive a SACK?  Normally RTO.Min >> max SACK delay,
so a retransmit means that either a packet was lost
or the SACK was lost.

I don't understand what problem you are seeing...

--brian

> Since the rfc (2960) requires you to SACK immediately on receiving
> the first DATA chunk, a similar rule could help in case of receiving
> retransmitted DATA chunks i.e. on receiving a DATA chunk that fills a
> hole and advances the cumulative TSN Ack.  Or,  is this left to the
> implementation ?
> Thanks!
> 
> Venkat
> 
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Wed Apr  3 05:31:05 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA28515
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 05:31:05 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA10573
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 05:31:08 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA09019;
	Wed, 3 Apr 2002 05:06:39 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA08988
	for <tsvwg@ns.ietf.org>; Wed, 3 Apr 2002 05:06:36 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-193.dallas.net [209.44.42.193])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA28285
	for <tsvwg@ietf.org>; Wed, 3 Apr 2002 05:06:31 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id EAA22307;
	Wed, 3 Apr 2002 04:05:23 -0600
Date: Wed, 3 Apr 2002 04:05:23 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Jonathan Stone <jonathan@dsg.stanford.edu>
Cc: Phillip Conrad <conrad@acm.org>, tsvwg@ietf.org
Subject: Re: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt
Message-ID: <20020403040523.A21473@openss7.org>
Reply-To: bidulock@openss7.org
References: <20020330204959.A27290@openss7.org> <200203310316.g2V3GhP17721@Gregorio.Stanford.EDU>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <200203310316.g2V3GhP17721@Gregorio.Stanford.EDU>; from jonathan@DSG.Stanford.EDU on Sat, Mar 30, 2002 at 07:16:38PM -0800
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Jonathan,

Why do you want to restrict me from calculating the checksum
in a mathematically-equivalent fashion?  (I have no outboard
accelerators.)

--brian

On Sat, 30 Mar 2002, Jonathan Stone wrote:

> 
> Here's an excerpt, slightly edited from an off-list email.
> Apologies to those who (like Brian) have seen it twice.
> 
> 
> My concern here is that for high-end SCTP implementations
> (gigabit, 10gigabit, and upward), outboard accelerators will be
> commonplace. Outboard accelerators make the check less end-to-end.
> The available data suggests that errors -- e.g., bits flipped in
> address registers in DMA engines, due to transient hardware errors --
> is a signficant source of errors.  They really *do* happen in the
> field, and at worrisome rates: on the order of 1 in 10,000 overall,
> and *much* higher for specific broken hosts.
> 
> Specification of the CRC computation should consider the issue of
> failure in outboard hardware acceleration of what is supposed to be an
> end-to-end error check.[*]  My take is that such consideration falls
> squarely under the "potential for causing harm" clause of RFC2119,
> section 6, not under the interoperability clause.
> 
> 
> I think that explicitly granting complete and utter freedom is a very
> poor engineering choice. I much prefer the current wording.  The most
> ``weaking'' of that tight specification I'd live with is two clauses:
> 
>         1. a MUST compute "AS IF" (i.e., must be mathematically
>            equivalent to) the current text
> 
>         2. a SHOULD NOT compute in some other mathematically-equivalent
>            way, since the mathematically-equivalent alternatives raised
>            in discussion are weaker when (for example) faced with errors
>            in outboard accelerators.
> 
> The argument I've seen presented against the MUST in the current
> draft, seems implicitly assumes that all computation is being done
> inside the end-host -- that is, using the end-host CPU and end-host
> buffers. That may be a valid assumption from the perspective of a
> pure-software implementation and implementor. But making that
> assumption misses the point behind the words, which is aimed at
> implementations where the ``end-to-end'' check is being done outboard.
> (If you like, consider Michael Tuexen's source code again, with the
> assumption that the address-manipulation is being done in outboard
> hardware, where the address arithmetic may  occasionally mess up.)
> 
> 
> I don't find applications like pretty-printing packet traces a
> convincing case. The pretty-printer can easily save the relevant 32
> bits before the CRC-check, and then restore them afterward.  It's not
> like packet pretty-printing is an actual protocol implementation :).

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Wed Apr  3 05:34:20 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA28577
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 05:34:19 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA10713
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 05:34:23 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA09360;
	Wed, 3 Apr 2002 05:11:31 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA09330
	for <tsvwg@ns.ietf.org>; Wed, 3 Apr 2002 05:11:29 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-193.dallas.net [209.44.42.193])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA28319
	for <tsvwg@ietf.org>; Wed, 3 Apr 2002 05:11:23 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id EAA22400;
	Wed, 3 Apr 2002 04:11:21 -0600
Date: Wed, 3 Apr 2002 04:11:20 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Randall Stewart <randall@stewart.chicago.il.us>
Cc: TSV WG list <tsvwg@ietf.org>
Subject: Re: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt]
Message-ID: <20020403041120.B21473@openss7.org>
Reply-To: bidulock@openss7.org
References: <3CA8EFB3.BC083B30@cisco.com> <3CA8F0AD.1F864E52@stewart.chicago.il.us>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <3CA8F0AD.1F864E52@stewart.chicago.il.us>; from randall@stewart.chicago.il.us on Mon, Apr 01, 2002 at 05:43:41PM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Randall,

Please see comments below...

On Mon, 01 Apr 2002, Randall Stewart wrote:

> Hmm..
> 
> Since this did not go to the working group in my reply-all
> and my second forward also failed (I think due to anti-spamming)
> I will retry.... forwarding to my home email first :>
> 
> More comments on the wonderful sctpcsum issue :>
> 
> R
> 
> 
> Randall Stewart at work wrote:
> > 
> > I have just looked over the whole thread... and I see both points
> > of view on this matter... I must agree Jonathan's arguments are (as
> > usual)
> > quite persuasive.
> > 
> > Now another thought I had ...  Where did this wording come from?
> > 
> > So I went back to RFC2960 and what do I find?
> > 
> > "
> >    When an SCTP packet is received, the receiver MUST first check the
> >    Adler-32 checksum:
> > 
> >    1) Store the received Adler-32 checksum value aside,
> > 
> > Stewart, et al.             Standards Track                    [Page 82]
> > 
> > RFC 2960          Stream Control Transmission Protocol      October 2000
> > 
> >    2) Replace the 32 bits of the checksum field in the received SCTP
> >       packet with all '0's and calculate an Adler-32 checksum value of
> >       the whole received packet.  And,
> > 
> >    3) Verify that the calculated Adler-32 checksum is the same as the
> >       received Adler-32 checksum.  If not, the receiver MUST treat the
> >       packet as an invalid SCTP packet.
> > 
> >    The default procedure for handling invalid SCTP packets is to
> >    silently discard them.
> > "
> > 
> > Looks kind of familiar does it not?
> > 
> > Now I think (if I have read correctly) what Jonathan has proposed is
> > something along
> > the lines of:
> > 
> > "
> >     After the packet is constructed (containing the SCTP common header and
> >     one or more control or DATA chunks), the transmitter MUST do the
> >     following:
> > 
> >     1) Fill in the proper Verification Tag in the SCTP common header and
> >        initialize the Checksum field to 0's.

Do you see the problem with this text????  Why does 1) say to fill in
the proper verification tag and initialize the checksum to 0 when the
paragraph aboves says that the packet already contains an SCTP common
header??  What was in the SCTP common header to begin with?  Random bits?

> > 
> >     2) Calculate the CRC-32c of the whole packet, including the SCTP common
> >        header and all the chunks.
> > 
> >     3) Put the resultant value into the Checksum field in the common header,
> >        and leave the rest of the bits unchanged.
> > 
> >     When an SCTP packet is received, the receiver MUST perform the checksum
> >     algorithm. It SHOULD perform this algorithm using the following procedure:
> > 
> >     1) Store the received CRC-32c value,
> > 
> >     2) Replace the 32 bits of the Checksum field in the received SCTP packet
> >        with all '0's and calculate a CRC-32c value of the whole received
> >        packet.  And,
> > 
> >     3) Verify that the calculated CRC-32c value is the same as the received
> >        CRC-32c value.  If not, the receiver MUST treat the packet as an
> >        invalid SCTP packet.
> > 

As Johnathan has not substantiate any of his statements regarding the
exact procedure after several requests for explanation, I propose that
you strike the following sentence and go with the modified text above.

> >     A fully conformant SCTP implementation SHOULD use this exact procedure.

There is no basis for this recommendation.  Another method of calculation
has not been shown to endanger anything, leaving this as an inappropriate
use of SHOULD.  Why use "fully conformant" to try an make the SHOULD a MUST?

--brian


> > "
> > 
> > Jonathan: does this capture your AS IF <current> spec sentiment?
> > 
> > Or did I miss it?
> > 
> > Would this be acceptable to all concerned? Or am I off base here?
> > 
> > R
> > --
> > Randall R. Stewart
> > ITD
> > Cisco Systems Inc.
> > rrs@cisco.com 815-342-5222(cell) or 815-477-2127(office)
> 
> 
> Now... I think there are two ways to go.. if this captured
> Jonathans idea and the working group agrees... we insert
> the above... Or just leave the text as is... I am fine
> with either method :>
> 
> I have -05 ready with leaving things the way they are...
> 
> Comments most welcome :-)
> 
> R
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Wed Apr  3 08:25:29 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA02532
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 08:25:29 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA23283
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 08:25:32 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA21378;
	Wed, 3 Apr 2002 07:56:47 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA21322
	for <tsvwg@ns.ietf.org>; Wed, 3 Apr 2002 07:56:44 -0500 (EST)
Received: from stewart.chicago.il.us ([64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA01434
	for <tsvwg@ietf.org>; Wed, 3 Apr 2002 07:56:39 -0500 (EST)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g33Cu3o79344;
	Wed, 3 Apr 2002 06:56:03 -0600 (CST)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CAAFBE3.FF590E7D@stewart.chicago.il.us>
Date: Wed, 03 Apr 2002 06:56:03 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: bidulock@openss7.org
CC: TSV WG list <tsvwg@ietf.org>
Subject: Re: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt]
References: <3CA8EFB3.BC083B30@cisco.com> <3CA8F0AD.1F864E52@stewart.chicago.il.us> <20020403041120.B21473@openss7.org>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

"Brian F. G. Bidulock" wrote:
> > Randall Stewart at work wrote:
> > > "
> > >     After the packet is constructed (containing the SCTP common header and
> > >     one or more control or DATA chunks), the transmitter MUST do the
> > >     following:
> > >
> > >     1) Fill in the proper Verification Tag in the SCTP common header and
> > >        initialize the Checksum field to 0's.
> 
> Do you see the problem with this text????  Why does 1) say to fill in
> the proper verification tag and initialize the checksum to 0 when the
> paragraph aboves says that the packet already contains an SCTP common
> header??  What was in the SCTP common header to begin with?  Random bits?
> 

Well lets see if we do NOT have a MUST here I could fill it with
anything I like 32, 64 or maybe 5,260  and still be completely compliant
with the wording. And yes even randon bits would be ok in the common
header if the MUST was not specified...


> > >
> > >     2) Calculate the CRC-32c of the whole packet, including the SCTP common
> > >        header and all the chunks.
> > >
> > >     3) Put the resultant value into the Checksum field in the common header,
> > >        and leave the rest of the bits unchanged.
> > >
> > >     When an SCTP packet is received, the receiver MUST perform the checksum
> > >     algorithm. It SHOULD perform this algorithm using the following procedure:
> > >
> > >     1) Store the received CRC-32c value,
> > >
> > >     2) Replace the 32 bits of the Checksum field in the received SCTP packet
> > >        with all '0's and calculate a CRC-32c value of the whole received
> > >        packet.  And,
> > >
> > >     3) Verify that the calculated CRC-32c value is the same as the received
> > >        CRC-32c value.  If not, the receiver MUST treat the packet as an
> > >        invalid SCTP packet.
> > >
> 
> As Johnathan has not substantiate any of his statements regarding the
> exact procedure after several requests for explanation, I propose that
> you strike the following sentence and go with the modified text above.
> 

Well, so far I have not seen an overwhelming consensus for
change... I will leave this to Scott and Allision to judge. I have
submitted an -05 version with typo's fixed but this left the same
i.e. not the above text that you commented on but the original.

I just don't see it as a problem Phil as convinced me off-line that with
a bit of sideways looking at it even the original wording will allow
Michaels code...

R
-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Wed Apr  3 09:23:20 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA04056
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 09:23:20 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA26766
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 09:23:22 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA25158;
	Wed, 3 Apr 2002 08:57:37 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA25128
	for <tsvwg@ns.ietf.org>; Wed, 3 Apr 2002 08:57:31 -0500 (EST)
Received: from mgw-x1.nokia.com (mgw-x1.nokia.com [131.228.20.21])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA03290
	for <tsvwg@ietf.org>; Wed, 3 Apr 2002 08:57:26 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir05nok.ntc.nokia.com (esvir05nokt.ntc.nokia.com [172.21.143.37])
	by mgw-x1.nokia.com (Switch-2.2.1/Switch-2.2.0) with ESMTP id g33Dueu11930
	for <tsvwg@ietf.org>; Wed, 3 Apr 2002 16:56:41 +0300 (EET DST)
Received: from esebh001.NOE.Nokia.com (unverified) by esvir05nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5a092dba1aac158f25077@esvir05nok.ntc.nokia.com>;
 Wed, 3 Apr 2002 16:57:28 +0300
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh001.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Wed, 3 Apr 2002 16:57:28 +0300
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt]
Date: Wed, 3 Apr 2002 16:57:27 +0300
Message-ID: <58AEDE27D48F884285007ABF827FF263257D4A@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt]
Thread-Index: AcHbD+OAT3f8hLpxSBSnc0oqMjsTYgABhPBg
To: <randall@stewart.chicago.il.us>, <bidulock@openss7.org>
Cc: <tsvwg@ietf.org>
X-OriginalArrivalTime: 03 Apr 2002 13:57:28.0164 (UTC) FILETIME=[7DA2DA40:01C1DB17]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id IAA25129
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi all!

	Please, see some comments below...

> > As Johnathan has not substantiate any of his statements 
> regarding the
> > exact procedure after several requests for explanation, I 
> propose that
> > you strike the following sentence and go with the modified 
> text above.
> > 
> 
> Well, so far I have not seen an overwhelming consensus for
> change... I will leave this to Scott and Allision to judge. I have
> submitted an -05 version with typo's fixed but this left the same
> i.e. not the above text that you commented on but the original.
> 
> I just don't see it as a problem Phil as convinced me 
> off-line that with
> a bit of sideways looking at it even the original wording will allow
> Michaels code...

	Why don't simply leaving things as they are and adding some note or something stating, for example: "Any other way of calculating the CRC that always leads to exactly the same result as the procedure stated above is also valid"?

	In my opinion, Brian made a good point. Especially when you are sending datagrams, you always know that those datagrams going to a specific association will always have the same port numbers, verification tag, and initial value of CRC (all 0s), i.e., same common header. Because of that, you can always skip the calculation of the CRC for the common header bytes since it will always lead to the same value (the only special cases are when the VT is not set to the "normal" value). When receiving a datagram, things are somehow more difficult to predict...

	So, in cases when you are only including one tiny chunk in the SCTP datagram (such as a SACK chunk without gaps or duplicate TSNs, for example, which in my opinion would be quite common), you would only need to calculate the CRC for the SACK bytes (20 bytes) instead of the whole datagram (32 bytes)... I don't know how much time this will save, but I don't see why you are not going to allow this as a legal practice...

	Comments?

	Thanks!

	BR Iván Arias Rodríguez

> R
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Wed Apr  3 13:56:34 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA12183
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 13:56:34 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA18377
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 13:56:24 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA17505;
	Wed, 3 Apr 2002 13:37:07 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA17466
	for <tsvwg@ns.ietf.org>; Wed, 3 Apr 2002 13:37:02 -0500 (EST)
Received: from c003.snv.cp.net (c003-h000.c003.snv.cp.net [209.228.32.214])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id NAA11751
	for <tsvwg@ietf.org>; Wed, 3 Apr 2002 13:36:59 -0500 (EST)
Received: (cpmta 21823 invoked from network); 3 Apr 2002 10:36:28 -0800
Received: from 66.123.21.234 (HELO littlejoy)
  by smtp.telocity.com (209.228.32.214) with SMTP; 3 Apr 2002 10:36:28 -0800
X-Sent: 3 Apr 2002 18:36:28 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: <Ivan.Arias-Rodriguez@nokia.com>, <randall@stewart.chicago.il.us>,
        <bidulock@openss7.org>
Cc: <tsvwg@ietf.org>
Subject: RE: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt]
Date: Wed, 3 Apr 2002 10:36:26 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJEEMMDAAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0)
In-Reply-To: <58AEDE27D48F884285007ABF827FF263257D4A@esebe017.NOE.Nokia.com>
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Ivan,

The draft makes no statement about the packet being in contagious memory nor
that the calculation be done simultaneously across the entire packet.  It
provides only the sequence of the calculation.  Once you take that
perspective, then Michael Tuexen and Brian Bidulock's schemes can be done
without altering procedures described within this draft.

Doug

On April 3, 2002 5:57 AM Ivan Arias-Rodriguez
(Ivan.Arias-Rodriguez@nokia.com) wrote:
>
> 	Hi all!
>
> 	Please, see some comments below...
>
> > > As Johnathan has not substantiate any of his statements
> > regarding the
> > > exact procedure after several requests for explanation, I
> > propose that
> > > you strike the following sentence and go with the modified
> > text above.
> > >
> >
> > Well, so far I have not seen an overwhelming consensus for
> > change... I will leave this to Scott and Allision to judge. I have
> > submitted an -05 version with typo's fixed but this left the same
> > i.e. not the above text that you commented on but the original.
> >
> > I just don't see it as a problem Phil as convinced me
> > off-line that with
> > a bit of sideways looking at it even the original wording will allow
> > Michaels code...
>
> 	Why don't simply leaving things as they are and adding some
> note or something stating, for example: "Any other way of
> calculating the CRC that always leads to exactly the same result
> as the procedure stated above is also valid"?
>
> 	In my opinion, Brian made a good point. Especially when you
> are sending datagrams, you always know that those datagrams going
> to a specific association will always have the same port numbers,
> verification tag, and initial value of CRC (all 0s), i.e., same
> common header. Because of that, you can always skip the
> calculation of the CRC for the common header bytes since it will
> always lead to the same value (the only special cases are when
> the VT is not set to the "normal" value). When receiving a
> datagram, things are somehow more difficult to predict...
>
> 	So, in cases when you are only including one tiny chunk in
> the SCTP datagram (such as a SACK chunk without gaps or duplicate
> TSNs, for example, which in my opinion would be quite common),
> you would only need to calculate the CRC for the SACK bytes (20
> bytes) instead of the whole datagram (32 bytes)... I don't know
> how much time this will save, but I don't see why you are not
> going to allow this as a legal practice...
>
> 	Comments?
>
> 	Thanks!
>
> 	BR Iván Arias Rodríguez
>
> > R
> > --
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Wed Apr  3 14:05:57 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA12450
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 14:05:57 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA19447
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 14:05:59 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA17770;
	Wed, 3 Apr 2002 13:44:47 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA17738
	for <tsvwg@ns.ietf.org>; Wed, 3 Apr 2002 13:44:44 -0500 (EST)
Received: from ihemail1.firewall.lucent.com (ihemail1.lucent.com [192.11.222.161])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA11900
	for <tsvwg@ietf.org>; Wed, 3 Apr 2002 13:44:41 -0500 (EST)
Received: from ihrh1.emsr.lucent.com (h135-1-218-53.lucent.com [135.1.218.53])
	by ihemail1.firewall.lucent.com (Switch-2.1.3/Switch-2.1.0) with ESMTP id g33IigX23107;
	Wed, 3 Apr 2002 13:44:42 -0500 (EST)
Received: from new-wopr.eng.ascend.com by ihrh1.emsr.lucent.com (8.8.8+Sun/EMS-1.5 Solaris/emsr)
	id MAA19270 for ; Wed, 3 Apr 2002 12:44:39 -0600 (CST)
Received: from grigri.eng.ascend.com (grigri.eng.ascend.com [135.140.53.45])
	by new-wopr.eng.ascend.com (8.10.2+Sun/8.10.2) with ESMTP id g33IifL28328;
	Wed, 3 Apr 2002 10:44:41 -0800 (PST)
Received: from lucent.com by grigri.eng.ascend.com (8.8.8+Sun/SMI-SVR4)
	id KAA24190; Wed, 3 Apr 2002 10:44:40 -0800 (PST)
Message-ID: <3CAB4E25.9E572AD2@lucent.com>
Date: Wed, 03 Apr 2002 10:47:01 -0800
From: Jacob Heitz <jheitz@lucent.com>
Organization: Lucent Technologies, Internetworking Systems, Software Engineering, 
 Alameda
X-Mailer: Mozilla 4.77 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Douglas Otis <dotis@sanlight.net>
CC: tsvwg@ietf.org
Subject: Re: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt]
References: <NEBBJGDMMLHHCIKHGBEJEEMMDAAA.dotis@sanlight.net>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Douglas Otis wrote:
> The draft makes no statement about the packet being in contagious memory
                                                         ^^^^^^^^^^
Does this mean it has a virus ? :-)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Wed Apr  3 14:24:54 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA13033
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 14:24:54 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA20411
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 14:24:57 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA18525;
	Wed, 3 Apr 2002 13:59:22 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA18497
	for <tsvwg@ns.ietf.org>; Wed, 3 Apr 2002 13:59:19 -0500 (EST)
Received: from c003.snv.cp.net (c003-h015.c003.snv.cp.net [209.228.32.229])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id NAA12246
	for <tsvwg@ietf.org>; Wed, 3 Apr 2002 13:59:16 -0500 (EST)
Received: (cpmta 27681 invoked from network); 3 Apr 2002 10:58:48 -0800
Received: from 66.123.21.234 (HELO littlejoy)
  by smtp.telocity.com (209.228.32.229) with SMTP; 3 Apr 2002 10:58:48 -0800
X-Sent: 3 Apr 2002 18:58:48 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: "Jacob Heitz" <jheitz@lucent.com>
Cc: <tsvwg@ietf.org>
Subject: RE: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt]
Date: Wed, 3 Apr 2002 10:58:46 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJOEMNDAAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0)
In-Reply-To: <3CAB4E25.9E572AD2@lucent.com>
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Jacob,

Opps, contiguous memory.

>
> Douglas Otis wrote:
> > The draft makes no statement about the packet being in contagious memory
>                                                          ^^^^^^^^^^
> Does this mean it has a virus ? :-)


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Wed Apr  3 14:30:36 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA13336
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 14:30:36 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA20801
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 14:30:38 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA19658;
	Wed, 3 Apr 2002 14:08:17 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA19629
	for <tsvwg@ns.ietf.org>; Wed, 3 Apr 2002 14:08:14 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-193.dallas.net [209.44.42.193])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA12504
	for <tsvwg@ietf.org>; Wed, 3 Apr 2002 14:08:10 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id NAA29084;
	Wed, 3 Apr 2002 13:07:31 -0600
Date: Wed, 3 Apr 2002 13:07:31 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Douglas Otis <dotis@sanlight.net>
Cc: Ivan.Arias-Rodriguez@nokia.com, randall@stewart.chicago.il.us,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt]
Message-ID: <20020403130731.A28856@openss7.org>
Reply-To: bidulock@openss7.org
References: <58AEDE27D48F884285007ABF827FF263257D4A@esebe017.NOE.Nokia.com> <NEBBJGDMMLHHCIKHGBEJEEMMDAAA.dotis@sanlight.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <NEBBJGDMMLHHCIKHGBEJEEMMDAAA.dotis@sanlight.net>; from dotis@sanlight.net on Wed, Apr 03, 2002 at 10:36:26AM -0800
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Douglas,

Then why the objection to mathematically equivalent procedures?  If
the draft clearly stated your interpretation: that the MUST applies to
the "calculation" and not the "method," there would not be a problem.

On the contrary, others have expressed that this interpretation is
not true.  Which is it?  Can we clarify the text (perhaps with a note)
so that one interpretation can be taken.

--brian


On Wed, 03 Apr 2002, Douglas Otis wrote:

> Ivan,
> 
> The draft makes no statement about the packet being in contagious memory nor
> that the calculation be done simultaneously across the entire packet.  It
> provides only the sequence of the calculation.  Once you take that
> perspective, then Michael Tuexen and Brian Bidulock's schemes can be done
> without altering procedures described within this draft.
> 
> Doug
> 
> On April 3, 2002 5:57 AM Ivan Arias-Rodriguez
> (Ivan.Arias-Rodriguez@nokia.com) wrote:
> >
> > 	Hi all!
> >
> > 	Please, see some comments below...
> >
> > > > As Johnathan has not substantiate any of his statements
> > > regarding the
> > > > exact procedure after several requests for explanation, I
> > > propose that
> > > > you strike the following sentence and go with the modified
> > > text above.
> > > >
> > >
> > > Well, so far I have not seen an overwhelming consensus for
> > > change... I will leave this to Scott and Allision to judge. I have
> > > submitted an -05 version with typo's fixed but this left the same
> > > i.e. not the above text that you commented on but the original.
> > >
> > > I just don't see it as a problem Phil as convinced me
> > > off-line that with
> > > a bit of sideways looking at it even the original wording will allow
> > > Michaels code...
> >
> > 	Why don't simply leaving things as they are and adding some
> > note or something stating, for example: "Any other way of
> > calculating the CRC that always leads to exactly the same result
> > as the procedure stated above is also valid"?
> >
> > 	In my opinion, Brian made a good point. Especially when you
> > are sending datagrams, you always know that those datagrams going
> > to a specific association will always have the same port numbers,
> > verification tag, and initial value of CRC (all 0s), i.e., same
> > common header. Because of that, you can always skip the
> > calculation of the CRC for the common header bytes since it will
> > always lead to the same value (the only special cases are when
> > the VT is not set to the "normal" value). When receiving a
> > datagram, things are somehow more difficult to predict...
> >
> > 	So, in cases when you are only including one tiny chunk in
> > the SCTP datagram (such as a SACK chunk without gaps or duplicate
> > TSNs, for example, which in my opinion would be quite common),
> > you would only need to calculate the CRC for the SACK bytes (20
> > bytes) instead of the whole datagram (32 bytes)... I don't know
> > how much time this will save, but I don't see why you are not
> > going to allow this as a legal practice...
> >
> > 	Comments?
> >
> > 	Thanks!
> >
> > 	BR Iván Arias Rodríguez
> >
> > > R
> > > --
> > > Randall R. Stewart
> > > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Wed Apr  3 14:57:16 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA14020
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 14:57:16 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA22428
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 14:57:19 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA20861;
	Wed, 3 Apr 2002 14:30:50 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA20824
	for <tsvwg@ns.ietf.org>; Wed, 3 Apr 2002 14:30:46 -0500 (EST)
Received: from c003.snv.cp.net (c003-h004.c003.snv.cp.net [209.228.32.218])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id OAA13347
	for <tsvwg@ietf.org>; Wed, 3 Apr 2002 14:30:42 -0500 (EST)
Received: (cpmta 11137 invoked from network); 3 Apr 2002 11:30:09 -0800
Received: from 66.123.21.234 (HELO littlejoy)
  by smtp.telocity.com (209.228.32.218) with SMTP; 3 Apr 2002 11:30:09 -0800
X-Sent: 3 Apr 2002 19:30:09 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: <bidulock@openss7.org>
Cc: <Ivan.Arias-Rodriguez@nokia.com>, <randall@stewart.chicago.il.us>,
        <tsvwg@ietf.org>
Subject: RE: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt]
Date: Wed, 3 Apr 2002 11:30:08 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJGEMPDAAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0)
In-Reply-To: <20020403130731.A28856@openss7.org>
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Brian,

Is it within the draft or Jonathan's statements you are objecting to?

The draft already indicates optional methods to calculate the CRC as
explanation with regard to initialization, as example.

The draft also indicates that the code is informative.  This leaves an
abstract calculation to be derived and implemented in whatever fashion
desired.  The draft also indicates there are many ways this can be done.
Beyond inventing a new algorithm, what is excluded by the draft?

Doug

On April 3, 2002 11:08 AM Brian F. G. Bidulock (bidulock@openss7.org) wrote:
> Douglas,
>
> Then why the objection to mathematically equivalent procedures?  If
> the draft clearly stated your interpretation: that the MUST applies to
> the "calculation" and not the "method," there would not be a problem.
>
> On the contrary, others have expressed that this interpretation is
> not true.  Which is it?  Can we clarify the text (perhaps with a note)
> so that one interpretation can be taken.
>
> --brian
>
>
> On Wed, 03 Apr 2002, Douglas Otis wrote:
>
> > Ivan,
> >
> > The draft makes no statement about the packet being in
> > contiguous memory nor that the calculation be done simultaneously
> > across the entire packet.  It provides only the sequence of the
> > calculation.  Once you take that perspective, then Michael Tuexen
> > and Brian Bidulock's schemes can be done without altering procedures
> > described within this draft.
> >
> > Doug
> >
> > On April 3, 2002 5:57 AM Ivan Arias-Rodriguez
> > (Ivan.Arias-Rodriguez@nokia.com) wrote:
> > >
> > > 	Hi all!
> > >
> > > 	Please, see some comments below...
> > >
> > > > > As Johnathan has not substantiate any of his statements
> > > > regarding the
> > > > > exact procedure after several requests for explanation, I
> > > > propose that
> > > > > you strike the following sentence and go with the modified
> > > > text above.
> > > > >
> > > >
> > > > Well, so far I have not seen an overwhelming consensus for
> > > > change... I will leave this to Scott and Allision to judge. I have
> > > > submitted an -05 version with typo's fixed but this left the same
> > > > i.e. not the above text that you commented on but the original.
> > > >
> > > > I just don't see it as a problem Phil as convinced me
> > > > off-line that with
> > > > a bit of sideways looking at it even the original wording will allow
> > > > Michaels code...
> > >
> > > 	Why don't simply leaving things as they are and adding some
> > > note or something stating, for example: "Any other way of
> > > calculating the CRC that always leads to exactly the same result
> > > as the procedure stated above is also valid"?
> > >
> > > 	In my opinion, Brian made a good point. Especially when you
> > > are sending datagrams, you always know that those datagrams going
> > > to a specific association will always have the same port numbers,
> > > verification tag, and initial value of CRC (all 0s), i.e., same
> > > common header. Because of that, you can always skip the
> > > calculation of the CRC for the common header bytes since it will
> > > always lead to the same value (the only special cases are when
> > > the VT is not set to the "normal" value). When receiving a
> > > datagram, things are somehow more difficult to predict...
> > >
> > > 	So, in cases when you are only including one tiny chunk in
> > > the SCTP datagram (such as a SACK chunk without gaps or duplicate
> > > TSNs, for example, which in my opinion would be quite common),
> > > you would only need to calculate the CRC for the SACK bytes (20
> > > bytes) instead of the whole datagram (32 bytes)... I don't know
> > > how much time this will save, but I don't see why you are not
> > > going to allow this as a legal practice...
> > >
> > > 	Comments?
> > >
> > > 	Thanks!
> > >
> > > 	BR Iván Arias Rodríguez
> > >
> > > > R
> > > > --
> > > > Randall R. Stewart
> > > > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> >
>
> --
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦
>


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Wed Apr  3 15:13:05 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA14369
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 15:13:05 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA23516
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 15:13:07 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA21797;
	Wed, 3 Apr 2002 14:44:04 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA21766
	for <tsvwg@ns.ietf.org>; Wed, 3 Apr 2002 14:43:57 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-193.dallas.net [209.44.42.193])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA13784
	for <tsvwg@ietf.org>; Wed, 3 Apr 2002 14:43:53 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id NAA29605;
	Wed, 3 Apr 2002 13:43:31 -0600
Date: Wed, 3 Apr 2002 13:43:31 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Douglas Otis <dotis@sanlight.net>
Cc: Ivan.Arias-Rodriguez@nokia.com, randall@stewart.chicago.il.us,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt]
Message-ID: <20020403134331.A29566@openss7.org>
Reply-To: bidulock@openss7.org
References: <20020403130731.A28856@openss7.org> <NEBBJGDMMLHHCIKHGBEJGEMPDAAA.dotis@sanlight.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <NEBBJGDMMLHHCIKHGBEJGEMPDAAA.dotis@sanlight.net>; from dotis@sanlight.net on Wed, Apr 03, 2002 at 11:30:08AM -0800
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Douglas,

I have been told differently by Randall and Johnathan on this
list.  They say that the algorithm must exactly follow the
procedure laid out in the draft to conform to the MUST.

The examples are informative and do not create or relax any
requirements, and may even be in error.

So, Randall, Jonhnathan, does the existing text permit me to
implement the caluclation "in whatever fashion desired" as
interpreted by Doug?

--brian

On Wed, 03 Apr 2002, Douglas Otis wrote:

> Brian,
> 
> Is it within the draft or Jonathan's statements you are objecting to?
> 
> The draft already indicates optional methods to calculate the CRC as
> explanation with regard to initialization, as example.
> 
> The draft also indicates that the code is informative.  This leaves an
> abstract calculation to be derived and implemented in whatever fashion
> desired.  The draft also indicates there are many ways this can be done.
> Beyond inventing a new algorithm, what is excluded by the draft?
> 
> Doug
> 

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr  3 17:05:51 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA17260
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 17:05:50 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA01416
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 17:05:54 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA29121;
	Wed, 3 Apr 2002 16:40:50 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA08658
	for <tsvwg@ns.ietf.org>; Tue, 2 Apr 2002 14:48:47 -0500 (EST)
Received: from seraph3.grc.nasa.gov (firewall-user@seraph3.lerc.nasa.gov [128.156.10.12])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA29886
	for <tsvwg@ietf.org>; Tue, 2 Apr 2002 14:48:45 -0500 (EST)
Received: from lombok-fi.lerc.nasa.gov (lombok-fi.lerc.nasa.gov [139.88.112.33])
	by seraph3.grc.nasa.gov (Postfix) with ESMTP id C0B3F640FB
	for <tsvwg@ietf.org>; Tue,  2 Apr 2002 14:46:16 -0500 (EST)
Received: from guns.lerc.nasa.gov (guns.lerc.nasa.gov [139.88.87.35])
	by lombok-fi.lerc.nasa.gov (NASA LeRC 8.9.1.1/8.9.1) with ESMTP id OAA12004;
	Tue, 2 Apr 2002 14:46:16 -0500 (EST)
Received: from guns.lerc.nasa.gov (localhost.lerc.nasa.gov [127.0.0.1]) by guns.lerc.nasa.gov with ESMTP (NASA LeRC 8.7.4.1/2.01-local)
        id OAA77759; Tue, 2 Apr 2002 14:46:10 -0500 (EST)
Message-Id: <200204021946.OAA77759@guns.lerc.nasa.gov>
To: Randall Stewart <randall@stewart.chicago.il.us>
From: Mark Allman <mallman@grc.nasa.gov>
Reply-To: mallman@grc.nasa.gov
Cc: venkat venkatsubra <venkats@austin.ibm.com>,
        Ivan.Arias-Rodriguez@nokia.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] delayed SACK when sender is retransmitting 
Organization: BBN Technologies/NASA GRC
Song-of-the-Day: Just What I Needed
Date: Tue, 02 Apr 2002 14:46:10 -0500
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


> Mark thinks its ok to SACK right away IF and only IF you have a
> gap that is being filled in.

Almost.

Mark thinks its OK to SACK immediatly if we are filling in all or
part of a gap **or** the incoming segment creates or enlarges a gap.

From RFC 2581:

   Out-of-order data segments SHOULD be acknowledged immediately, in
   order to accelerate loss recovery.  To trigger the fast retransmit
   algorithm, the receiver SHOULD send an immediate duplicate ACK when
   it receives a data segment above a gap in the sequence space.  To
   provide feedback to senders recovering from losses, the receiver
   SHOULD send an immediate ACK when it receives a data segment that
   fills in all or part of a gap in the sequence space.

allman


--
Mark Allman -- BBN/NASA GRC -- http://roland.grc.nasa.gov/~mallman/


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr  3 17:05:51 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA17273
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 17:05:51 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA01429
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 17:05:55 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA29225;
	Wed, 3 Apr 2002 16:41:08 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA27845
	for <tsvwg@optimus.ietf.org>; Wed, 3 Apr 2002 16:28:14 -0500 (EST)
Received: from smtp.comcast.net (smtp.comcast.net [24.153.64.2])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA16283
	for <tsvwg@ietf.org>; Wed, 3 Apr 2002 16:28:10 -0500 (EST)
Received: from livingroom.acm.org
 (pcp01480838pcs.ncstl01.de.comcast.net [68.82.70.12])
 by mtaout04.icomcast.net (iPlanet Messaging Server 5.1 (built Feb  6 2002))
 with ESMTP id <0GU000GPPGY7S9@mtaout04.icomcast.net> for tsvwg@ietf.org; Wed,
 03 Apr 2002 16:27:43 -0500 (EST)
Date: Wed, 03 Apr 2002 15:17:08 -0500
From: Phillip Conrad <conrad@acm.org>
Subject: Re: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt]
In-reply-to: <3CAB4E25.9E572AD2@lucent.com>
X-Sender: phillipconrad@mail.comcast.net
To: tsvwg@ietf.org
Message-id: <5.1.0.14.2.20020403150853.03937c80@joda.cis.temple.edu>
MIME-version: 1.0
X-Mailer: QUALCOMM Windows Eudora Version 5.1
Content-type: text/plain; charset=us-ascii; format=flowed
Content-transfer-encoding: 7BIT
References: <NEBBJGDMMLHHCIKHGBEJEEMMDAAA.dotis@sanlight.net>
Content-Transfer-Encoding: 7BIT
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7BIT

At 10:47 AM 4/3/2002 -0800, Jacob Heitz wrote:
>Douglas Otis wrote:
> > The draft makes no statement about the packet being in contagious memory
>                                                          ^^^^^^^^^^
>Does this mean it has a virus ? :-)

I suspect that Doug may have intended to say "contiguous" rather than 
"contagious".

Merriam-Webster's defines:
  contiguous: being in actual contact : touching along a boundary or at a point
  contagious: communicable by contact : CATCHING <contagious diseases>.

But you never know: those physicists and engineers are developing new types
of memory all the time.

On a more serious note, I've given this more thought and discussed it with
some folks.

All in all, unless there are NEW voices to the contrary, or NEW points
raised, I'm convinced that the issue raised by Brian (that the current text
unnecessarily/inappropriately constrains the implementor) while having
enough merit to warrant the discussion that has already taken place, does in
the end does not have enough merit to warrant any change in the current text.

The two important things that the draft needs to capture are that

(1) a checksum calculation MUST be done on incoming packets, and that

(2) the definition of this required checksum calculation is "CRC-32c over
the received packet, but with the checksum field in that packet replaced by
zeros."

I think the current language does this just fine.    The main objection to
the current language is that it inappropriately constrains the implementor.
For a brief time, I was persuaded by this argument, but I am no longer so
persuaded, and would suggest that the current language, unmodified, is fine.

 From what I've seen so far, reluctance to change the current language stems
from any or all of three reasons:

(1) the *particular* language to which Brian Bidulock objects was already
*previously* approved at every level of the IETF review process; as evidence
of this, see RFC2960, Section 8.3, pages 81-82, where the same language
appears, with the only difference being "CRC32c" instead of "Adler-32?.
This doesn't PROVE that the language is appropriate, since no-one would
claim that the IETF process is "infallible", but it does provide a strong
case that this language has already undergone significant review, and that
the rough consensus was "thumbs up".

(2) There is concern that changing the language may open up what is
currently perceived by some to be a water-tight document, to all kinds of
philistine and libertine practices such as "not computing the checksum at
all", or computing it on an off-board processor
or out-board motor, or in an off-shore betting establishment.

Personally, I'm not convinced that anything in the *current* language, or
for that matter any of the currently proposed alternatives, is going to have
much impact one way or the other on what an implementer does in regard to
"how" the checksum is calculated.      If it is really very important to
prohibit or discourage off-board processing of the checksum (and I'm now
convinced by Jonathan and others that perhaps we should) then I think we
should  *say that explicitly*.

Bottom line, these ARE internal implementation details, not visible on the
wire unless the hardware actually fails (unlike choosing explicitly to NOT
verify the checksum at all, which is easy to detect on the wire).   But, to
the extent that colleagues who are more knowledgeable about checksums (e.g.
Jonathan Stone) see this issue differently, I'm more than willing to defer
to their judgement about the power of the current language to discourage the
practice of "unsafe checksumming."

(3) The only counter-example of how one MIGHT want to compute the checksum
differently from the straightforward way described in the current draft is
the one in Michael Tuexen's ethereal code (previously posted on the list).
    However, on further consideration, I think this can be seen to be in
compliance with the current draft.    I'll leave the details of that to a
postscript.

All in all, unless there are NEW voices to the contrary, or NEW points
raised, I'm convinced that the issue raised by Brian (that the current text
unnecessarily/inappropriately constrains the implementor) while having
enough merit to warrant the discussion that has already taken place, does in
the end does not have enough merit to warrant any change in the current text.

Cheers,
Phill


PS: Two appendices...  first, how the current draft can permit Michael
Tuxen's code, and second, a few comments from Lewis Carroll that may be in
order here...

"1) Store the received CRC-32c value".

The value is most certainly stored by Michael's code (albeit in its original
place.)

"2) Replace the 32 bits of the Checksum field in the received SCTP packet
with all '0's and calculate a CRC-32c value of the whole received packet."

The word "replacement" here does not spell out whether it
is a physical movement of data, or whether it is a mathematical replacement.

The most straightforward interpretation  is a physical movement of data.
Michael's code, instead, does a mathematical replacement, by passing zeros
to the CRC32c function, bypassing the four bytes of the checksum field.

Also, note that "do X and do Y" doesn't necessarily mean "do X first, and do
Y second".   The replacement does not HAVE to PRECEDE the checksum
calculation; Michael's code integrates them into one step.   Perfectly fine.

"3) Verify that the calculated CRC-32c value is the same as the received
CRC-32c value. If not, the receiver MUST treat the packet as an invalid SCTP
packet. "    Actually, Michael's code doesn't spell out how this is done,
but presumably, the value returned at the end would be compared to the value
that is *still* stored in place in the packet buffer.  So, again, no problems.

A little quote from "Through the Looking Glass" by Lewis Caroll may be in
order here.

>`I don't know what you mean by "glory,"' Alice said.
>
>   Humpty Dumpty smiled contemptuously.  `Of course you don't--
>till I tell you.  I meant "there's a nice knock-down argument for
>you!"'
>
>   `But "glory" doesn't mean "a nice knock-down argument,"' Alice
>objected.
>
>   `When _I_ use a word,' Humpty Dumpty said in rather a scornful
>tone, `it means just what I choose it to mean--neither more nor
>less.'
>
>   `The question is,' said Alice, `whether you CAN make words mean
>so many different things.'
>
>   `The question is,' said Humpty Dumpty, `which is to be master--
>that's all.'

More such fun available at: http://www.textlibrary.com/download/thru-loo.txt



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr  3 17:06:16 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA17293
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 17:06:16 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA01461
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 17:06:20 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA29071;
	Wed, 3 Apr 2002 16:40:31 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA21327
	for <tsvwg@optimus.ietf.org>; Tue, 2 Apr 2002 11:05:38 -0500 (EST)
Received: from seraph2.grc.nasa.gov (firewall-user@seraph2.lerc.nasa.gov [128.156.10.11])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA20579
	for <tsvwg@ietf.org>; Tue, 2 Apr 2002 11:05:36 -0500 (EST)
Received: from lombok-fi.lerc.nasa.gov (lombok-fi.lerc.nasa.gov [139.88.112.33])
	by seraph2.grc.nasa.gov (Postfix) with ESMTP id 51B04C68FC
	for <tsvwg@ietf.org>; Tue,  2 Apr 2002 11:05:38 -0500 (EST)
Received: from guns.lerc.nasa.gov (guns.lerc.nasa.gov [139.88.87.35])
	by lombok-fi.lerc.nasa.gov (NASA LeRC 8.9.1.1/8.9.1) with ESMTP id LAA00971;
	Tue, 2 Apr 2002 11:05:37 -0500 (EST)
Received: from guns.lerc.nasa.gov (localhost.lerc.nasa.gov [127.0.0.1]) by guns.lerc.nasa.gov with ESMTP (NASA LeRC 8.7.4.1/2.01-local)
        id LAA21857; Tue, 2 Apr 2002 11:05:31 -0500 (EST)
Message-Id: <200204021605.LAA21857@guns.lerc.nasa.gov>
To: Randall Stewart <randall@stewart.chicago.il.us>
From: Mark Allman <mallman@grc.nasa.gov>
Reply-To: mallman@grc.nasa.gov
Cc: Arun Prasad <arun@netlab.hcltech.com>,
        venkat venkatsubra <venkats@austin.ibm.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] delayed SACK when sender is retransmitting 
Organization: BBN Technologies/NASA GRC
Song-of-the-Day: Just What I Needed
Date: Tue, 02 Apr 2002 11:05:31 -0500
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


> I think Arun has mad the correct observation here.
> 
> It is really impossible to tell the difference between
> a retransmission and a network delayed packet. 

Yeah, I agree.

But, I think the packet should be immediatly SACKed.  In RFC 2581 we
added some language that says you should immedaitly ACK packets that
either create a hole in the sequence (or TSN, in this case) space or
that fill in the hole (or part of a hole).  Doing so provides the
quickest possible feedback to the sender as to the state of the
world -- which is a good thing for loss recovery algorithms.

allman


--
Mark Allman -- BBN/NASA GRC -- http://roland.grc.nasa.gov/~mallman/


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr  3 17:08:28 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA17359
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 17:08:27 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA01632
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 17:08:31 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA28988;
	Wed, 3 Apr 2002 16:39:47 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA28030
	for <tsvwg@optimus.ietf.org>; Mon, 1 Apr 2002 07:16:30 -0500 (EST)
Received: from CNRI.Reston.VA.US (localhost [127.0.0.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA22036;
	Mon, 1 Apr 2002 07:16:06 -0500 (EST)
Message-Id: <200204011216.HAA22036@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: IETF-Announce: ;
Cc: tsvwg@ietf.org
From: Internet-Drafts@ietf.org
Reply-to: Internet-Drafts@ietf.org
Date: Mon, 01 Apr 2002 07:16:06 -0500
Subject: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-initwin-02.txt
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the Transport Area Working Group Working Group of the IETF.

	Title		: Increasing TCP's Initial Window
	Author(s)	: M. Allman, S. Floyd, C. Partridge
	Filename	: draft-ietf-tsvwg-initwin-02.txt
	Pages		: 12
	Date		: 29-Mar-02
	
This document specifies an increase in the permitted initial window
for TCP from one segment to roughly 4K bytes.  This document also
discusses the advantages and disadvantages of the change, outlining
experimental results that indicate the costs and benefits.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-initwin-02.txt

To remove yourself from the IETF Announcement list, send a message to 
ietf-announce-request with the word unsubscribe in the body of the message.

Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-tsvwg-initwin-02.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-tsvwg-initwin-02.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<20020329130908.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-tsvwg-initwin-02.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-tsvwg-initwin-02.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<20020329130908.I-D@ietf.org>

--OtherAccess--

--NextPart--




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr  3 17:10:12 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA17407
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 17:10:12 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA01725
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 17:10:16 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA29032;
	Wed, 3 Apr 2002 16:40:06 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA23921
	for <tsvwg@ns.ietf.org>; Mon, 1 Apr 2002 12:13:32 -0500 (EST)
Received: from sj-msg-core-4.cisco.com (sj-msg-core-4.cisco.com [171.71.163.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA06198
	for <tsvwg@ietf.org>; Mon, 1 Apr 2002 12:13:29 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (IDENT:mirapoint@mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-4.cisco.com (8.12.2/8.12.2) with ESMTP id g31HD1I6029213
	for <tsvwg@ietf.org>; Mon, 1 Apr 2002 09:13:01 -0800 (PST)
Received: from cisco.com (ssh-sjc-1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id ABX36084;
	Mon, 1 Apr 2002 09:13:00 -0800 (PST)
Message-ID: <3CA8951B.DC94746F@cisco.com>
Date: Mon, 01 Apr 2002 11:12:59 -0600
From: Randall Stewart at work <rrs@cisco.com>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: tsvwg@ietf.org
Subject: Re: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt
References: <200203300410.g2U4AmC23317@newdev.harvard.edu> <3CA7A29E.B0676A86@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

All:

I just noticed that this did not go to the list...opps...

> 
> I have just looked over the whole thread... and I see both points
> of view on this matter... I must agree Jonathan's arguments are (as
> usual)
> quite persuasive.
> 
> Now another thought I had ...  Where did this wording come from?
> 
> So I went back to RFC2960 and what do I find?
> 
> "
>    When an SCTP packet is received, the receiver MUST first check the
>    Adler-32 checksum:
> 
>    1) Store the received Adler-32 checksum value aside,
> 
> Stewart, et al.             Standards Track                    [Page 82]
> 
> RFC 2960          Stream Control Transmission Protocol      October 2000
> 
>    2) Replace the 32 bits of the checksum field in the received SCTP
>       packet with all '0's and calculate an Adler-32 checksum value of
>       the whole received packet.  And,
> 
>    3) Verify that the calculated Adler-32 checksum is the same as the
>       received Adler-32 checksum.  If not, the receiver MUST treat the
>       packet as an invalid SCTP packet.
> 
>    The default procedure for handling invalid SCTP packets is to
>    silently discard them.
> "
> 
> Looks kind of familiar does it not?
> 
> Now I think (if I have read correctly) what Jonathan has proposed is
> something along the lines of:
> 
> "
>     After the packet is constructed (containing the SCTP common header and
>     one or more control or DATA chunks), the transmitter MUST do the
>     following:
> 
>     1) Fill in the proper Verification Tag in the SCTP common header and
>        initialize the Checksum field to 0's.
> 
>     2) Calculate the CRC-32c of the whole packet, including the SCTP common
>        header and all the chunks.
> 
>     3) Put the resultant value into the Checksum field in the common header,
>        and leave the rest of the bits unchanged.
> 
>     When an SCTP packet is received, the receiver MUST perform the checksum
>     algorithm. It SHOULD perform this algorithm using the following procedure:
> 
>     1) Store the received CRC-32c value,
> 
>     2) Replace the 32 bits of the Checksum field in the received SCTP packet
>        with all '0's and calculate a CRC-32c value of the whole received
>        packet.  And,
> 
>     3) Verify that the calculated CRC-32c value is the same as the received
>        CRC-32c value.  If not, the receiver MUST treat the packet as an
>        invalid SCTP packet.
> 
>     A fully conformant SCTP implementation SHOULD use this exact procedure.
> "
> 
> Jonathan: does this capture your AS IF <current> spec sentiment?
> 
> Or did I miss it?
> 
> Would this be acceptable to all concerned? 

The other choice (that is at least acceptable to me) is to leave
the text alone :>

What is the opinion of the WG??

R

-- 
Randall R. Stewart
ITD
Cisco Systems Inc.
rrs@cisco.com 815-342-5222(cell) or 815-477-2127(office)


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr  3 17:28:43 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA17729
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 17:28:38 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA02200
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 17:28:43 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA01375;
	Wed, 3 Apr 2002 17:05:05 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA01348
	for <tsvwg@optimus.ietf.org>; Wed, 3 Apr 2002 17:04:59 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-193.dallas.net [209.44.42.193])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA17225
	for <tsvwg@ietf.org>; Wed, 3 Apr 2002 17:04:40 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id QAA31389;
	Wed, 3 Apr 2002 16:04:35 -0600
Date: Wed, 3 Apr 2002 16:04:35 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Phillip Conrad <conrad@acm.org>
Cc: tsvwg@ietf.org
Subject: Re: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt]
Message-ID: <20020403160435.C29566@openss7.org>
Reply-To: bidulock@openss7.org
References: <NEBBJGDMMLHHCIKHGBEJEEMMDAAA.dotis@sanlight.net> <3CAB4E25.9E572AD2@lucent.com> <5.1.0.14.2.20020403150853.03937c80@joda.cis.temple.edu>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <5.1.0.14.2.20020403150853.03937c80@joda.cis.temple.edu>; from conrad@acm.org on Wed, Apr 03, 2002 at 03:17:08PM -0500
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Phillip,

Please see comments below...

On Wed, 03 Apr 2002, Phillip Conrad wrote:

> At 10:47 AM 4/3/2002 -0800, Jacob Heitz wrote:
> >Douglas Otis wrote:
> > > The draft makes no statement about the packet being in contagious memory
> >                                                          ^^^^^^^^^^
> >Does this mean it has a virus ? :-)
> 
> I suspect that Doug may have intended to say "contiguous" rather than 
> "contagious".
> 
> Merriam-Webster's defines:
>   contiguous: being in actual contact : touching along a boundary or at a point
>   contagious: communicable by contact : CATCHING <contagious diseases>.
> 
> But you never know: those physicists and engineers are developing new types
> of memory all the time.
> 
> On a more serious note, I've given this more thought and discussed it with
> some folks.
> 
> All in all, unless there are NEW voices to the contrary, or NEW points
> raised, I'm convinced that the issue raised by Brian (that the current text
> unnecessarily/inappropriately constrains the implementor) while having
> enough merit to warrant the discussion that has already taken place, does in
> the end does not have enough merit to warrant any change in the current text.
> 
> The two important things that the draft needs to capture are that
> 
> (1) a checksum calculation MUST be done on incoming packets, and that
> 
> (2) the definition of this required checksum calculation is "CRC-32c over
> the received packet, but with the checksum field in that packet replaced by
> zeros."
> 
> I think the current language does this just fine.    The main objection to
> the current language is that it inappropriately constrains the implementor.
> For a brief time, I was persuaded by this argument, but I am no longer so
> persuaded, and would suggest that the current language, unmodified, is fine.
> 
>  From what I've seen so far, reluctance to change the current language stems
> from any or all of three reasons:
> 
> (1) the *particular* language to which Brian Bidulock objects was already
> *previously* approved at every level of the IETF review process; as evidence
> of this, see RFC2960, Section 8.3, pages 81-82, where the same language
> appears, with the only difference being "CRC32c" instead of "Adler-32?.
> This doesn't PROVE that the language is appropriate, since no-one would
> claim that the IETF process is "infallible", but it does provide a strong
> case that this language has already undergone significant review, and that
> the rough consensus was "thumbs up".

Unfortunately it is not the same language.  RFC 2960 says:

"    When an SCTP packet is received, the receiver MUST first check the
"    Adler-32 checksum:
" 
"    1) Store the received Adler-32 checksum value aside,
" 
"    2) Replace the 32 bits of the checksum field in the received SCTP
"       packet with all '0's and calculate an Adler-32 checksum value of
"       the whole received packet.  And,
" 
"    3) Verify that the calculated Adler-32 checksum is the same as the
"       received Adler-32 checksum.  If not, the receiver MUST treat the
"       packet as an invalid SCTP packet.
" 
"    The default procedure for handling invalid SCTP packets is to
"    silently discard them.

CSUM-04 says:

"    When an SCTP packet is received, the receiver MUST first perform the 
"    following: 
"     
"    1) Store the received CRC-32c value, 
"     
"    2) Replace the 32 bits of the Checksum field in the received SCTP packet 
"       with all '0's and calculate a CRC-32c value of the whole received 
"       packet.  And, 
"     
"    3) Verify that the calculated CRC-32c value is the same as the received 
"       CRC-32c value.  If not, the receiver MUST treat the packet as an 
"       invalid SCTP packet. 
"     
"    The default procedure for handling invalid SCTP packets is to silently 
"    discard them. 

If we change:

"    When an SCTP packet is received, the receiver MUST first perform the 
"    following: 

to

"    When an SCTP packet is received, the receiver MUST first check the
"    CRC-32c checksum:

and make the CSUM text match the RFC 2960 text, I would have no problem.
This difference is one says that the check must be performed, the other
says how it must be performed.

> 
> (2) There is concern that changing the language may open up what is
> currently perceived by some to be a water-tight document, to all kinds of
> philistine and libertine practices such as "not computing the checksum at
> all", or computing it on an off-board processor
> or out-board motor, or in an off-shore betting establishment.
> 
> Personally, I'm not convinced that anything in the *current* language, or
> for that matter any of the currently proposed alternatives, is going to have
> much impact one way or the other on what an implementer does in regard to
> "how" the checksum is calculated.      If it is really very important to
> prohibit or discourage off-board processing of the checksum (and I'm now
> convinced by Jonathan and others that perhaps we should) then I think we
> should  *say that explicitly*.
> 
> Bottom line, these ARE internal implementation details, not visible on the
> wire unless the hardware actually fails (unlike choosing explicitly to NOT
> verify the checksum at all, which is easy to detect on the wire).   But, to
> the extent that colleagues who are more knowledgeable about checksums (e.g.
> Jonathan Stone) see this issue differently, I'm more than willing to defer
> to their judgement about the power of the current language to discourage the
> practice of "unsafe checksumming."

After reading the introduction, I would question that knowledge.

> 
> (3) The only counter-example of how one MIGHT want to compute the checksum
> differently from the straightforward way described in the current draft is
> the one in Michael Tuexen's ethereal code (previously posted on the list).
>     However, on further consideration, I think this can be seen to be in
> compliance with the current draft.    I'll leave the details of that to a
> postscript.

You have overlooked my two counterexamples: precalculating the SCTP
common header checksum on a per-association basis as a more efficient
starting point for the calculation, and the Van Jacobson 30-line send
approach of performing partial checksums on data while copying from the
user.  As CRC-32c is division, it can be performed as a series of partial
divisions which are later put together when the packet is bundled.  This
is done for TCP where software checksums are performed all the time.
To see how the approach is done, see any BSD NET4 based implementation
of TCP.

> 
> All in all, unless there are NEW voices to the contrary, or NEW points
> raised, I'm convinced that the issue raised by Brian (that the current text
> unnecessarily/inappropriately constrains the implementor) while having
> enough merit to warrant the discussion that has already taken place, does in
> the end does not have enough merit to warrant any change in the current text.

I agree: please do not change the text over RFC 2960: this means that
the text in the draft needs to be changed back from:

 <    When an SCTP packet is received, the receiver MUST first perform the 
 <    following: 

to read:

 >    When an SCTP packet is received, the receiver MUST first check the
 >    CRC-32c checksum:

--brian

> 
> Cheers,
> Phill
> 
> 
> PS: Two appendices...  first, how the current draft can permit Michael
> Tuxen's code, and second, a few comments from Lewis Carroll that may be in
> order here...
> 
> "1) Store the received CRC-32c value".
> 
> The value is most certainly stored by Michael's code (albeit in its original
> place.)
> 
> "2) Replace the 32 bits of the Checksum field in the received SCTP packet
> with all '0's and calculate a CRC-32c value of the whole received packet."
> 
> The word "replacement" here does not spell out whether it
> is a physical movement of data, or whether it is a mathematical replacement.

Johnathan objected to "mathematical equivalents".

> 
> The most straightforward interpretation  is a physical movement of data.
> Michael's code, instead, does a mathematical replacement, by passing zeros
> to the CRC32c function, bypassing the four bytes of the checksum field.
> 
> Also, note that "do X and do Y" doesn't necessarily mean "do X first, and do
> Y second".   The replacement does not HAVE to PRECEDE the checksum
> calculation; Michael's code integrates them into one step.   Perfectly fine.

Johnathan (and Randall) objected to reordering of the procedure.

> 
> "3) Verify that the calculated CRC-32c value is the same as the received
> CRC-32c value. If not, the receiver MUST treat the packet as an invalid SCTP
> packet. "    Actually, Michael's code doesn't spell out how this is done,
> but presumably, the value returned at the end would be compared to the value
> that is *still* stored in place in the packet buffer.  So, again, no problems.
> 
> A little quote from "Through the Looking Glass" by Lewis Caroll may be in
> order here.

Lewis Caroll was a mathematician and logician, and probably a poor
engineer.

--brian

> 
> >`I don't know what you mean by "glory,"' Alice said.
> >
> >   Humpty Dumpty smiled contemptuously.  `Of course you don't--
> >till I tell you.  I meant "there's a nice knock-down argument for
> >you!"'
> >
> >   `But "glory" doesn't mean "a nice knock-down argument,"' Alice
> >objected.
> >
> >   `When _I_ use a word,' Humpty Dumpty said in rather a scornful
> >tone, `it means just what I choose it to mean--neither more nor
> >less.'
> >
> >   `The question is,' said Alice, `whether you CAN make words mean
> >so many different things.'
> >
> >   `The question is,' said Humpty Dumpty, `which is to be master--
> >that's all.'
> 
> More such fun available at: http://www.textlibrary.com/download/thru-loo.txt
> 
> 
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr  3 18:39:07 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA19019
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 18:39:07 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id SAA06201
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 18:39:10 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA04298;
	Wed, 3 Apr 2002 18:09:28 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA04269
	for <tsvwg@optimus.ietf.org>; Wed, 3 Apr 2002 18:09:24 -0500 (EST)
Received: from c003.snv.cp.net (c003-h000.c003.snv.cp.net [209.228.32.214])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id SAA18381
	for <tsvwg@ietf.org>; Wed, 3 Apr 2002 18:09:21 -0500 (EST)
Received: (cpmta 11061 invoked from network); 3 Apr 2002 15:08:54 -0800
Received: from 66.123.21.234 (HELO littlejoy)
  by smtp.telocity.com (209.228.32.214) with SMTP; 3 Apr 2002 15:08:54 -0800
X-Sent: 3 Apr 2002 23:08:54 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: "Phillip Conrad" <conrad@acm.org>, <tsvwg@ietf.org>
Subject: RE: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt]
Date: Wed, 3 Apr 2002 15:08:53 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJEENIDAAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0)
In-Reply-To: <5.1.0.14.2.20020403150853.03937c80@joda.cis.temple.edu>
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Phillip,

I think you will find defining where transport CRC is preformed to be
problematic.  As we pursue ever higher throughputs, it is interesting to
note next generation IO solutions replace PCI by emulation with serial
packets also protected by CRC.  Although it would be impossible to trust
equipment beyond your control as would be the case with a MAN or WAN
environment, the quality of local equipment is within your control.
Interfaces where errors go unchecked is being removed in newer designs.

There is no protection from a stray cosmic ray from flipping a bit.  A bit
error in hardware may have greater ramifications but as there is never a
time where the CPU is not dependent upon a great deal of hardware.
Positioning the algorithm closer to the CPU may have advantages but these
advantages may be reduced by a more complex system that has greater
opportunity to go astray.

Doug


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr  3 18:47:24 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA19186
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 18:47:24 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id SAA06556
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 18:47:27 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA04590;
	Wed, 3 Apr 2002 18:19:22 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA04558
	for <tsvwg@optimus.ietf.org>; Wed, 3 Apr 2002 18:19:18 -0500 (EST)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA18484
	for <tsvwg@ietf.org>; Wed, 3 Apr 2002 18:19:14 -0500 (EST)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g33Mejo80471;
	Wed, 3 Apr 2002 16:40:45 -0600 (CST)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CAB84EC.6BB3226F@stewart.chicago.il.us>
Date: Wed, 03 Apr 2002 16:40:45 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: mallman@grc.nasa.gov
CC: venkat venkatsubra <venkats@austin.ibm.com>,
        Ivan.Arias-Rodriguez@nokia.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] delayed SACK when sender is retransmitting
References: <200204021946.OAA77759@guns.lerc.nasa.gov>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Mark:

I guess I was being a bit un-specific...

if a gap exists RFC2960 requires that you send a
SACK with no delay.. So the only case not covered that
you are stating needs to be SACK'd without delay is
when a TSN arrives and completes all outstanding
data. 

If there are still holes.. RFC2960 is going to make a
SACK go out right away.. so the only case not covered
is the "fills in the missing piece" case. 

R

Mark Allman wrote:
> 
> > Mark thinks its ok to SACK right away IF and only IF you have a
> > gap that is being filled in.
> 
> Almost.
> 
> Mark thinks its OK to SACK immediatly if we are filling in all or
> part of a gap **or** the incoming segment creates or enlarges a gap.
> 
> >From RFC 2581:
> 
>    Out-of-order data segments SHOULD be acknowledged immediately, in
>    order to accelerate loss recovery.  To trigger the fast retransmit
>    algorithm, the receiver SHOULD send an immediate duplicate ACK when
>    it receives a data segment above a gap in the sequence space.  To
>    provide feedback to senders recovering from losses, the receiver
>    SHOULD send an immediate ACK when it receives a data segment that
>    fills in all or part of a gap in the sequence space.
> 
> allman
> 
> --
> Mark Allman -- BBN/NASA GRC -- http://roland.grc.nasa.gov/~mallman/
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr  3 19:22:55 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA19774
	for <tsvwg-archive@odin.ietf.org>; Wed, 3 Apr 2002 19:22:54 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id TAA08560
	for tsvwg-archive@odin.ietf.org; Wed, 3 Apr 2002 19:22:56 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA07064;
	Wed, 3 Apr 2002 18:59:23 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA07029
	for <tsvwg@optimus.ietf.org>; Wed, 3 Apr 2002 18:59:19 -0500 (EST)
Received: from boreas.isi.edu (boreas.isi.edu [128.9.160.161])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA19406
	for <tsvwg@ietf.org>; Wed, 3 Apr 2002 18:59:17 -0500 (EST)
Received: from nit.isi.edu (nit.isi.edu [128.9.160.116])
	by boreas.isi.edu (8.11.6/8.11.2) with ESMTP id g33NxJT10803;
	Wed, 3 Apr 2002 15:59:19 -0800 (PST)
From: Aaron Falk <falk@ISI.EDU>
To: tsvwg@ietf.org
Cc: Carl Williams <carlw@docomolabs-usa.com>
Content-Type: text/plain
Content-Transfer-Encoding: 7bit
X-Mailer: Evolution/1.0.2 
Date: 03 Apr 2002 15:59:19 -0800
Message-Id: <1017878359.3859.206.camel@nit>
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] notes from L2 Triggers meeting
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

L2-Trigger unofficial meeting

Tuesday, March 18th, 2002
IETF-53 Minneapolis, MN

led by (& notes taken by)
Aaron Falk <falk@isi.edu> and
Carl Williams <carlw@docomolabs-usa.com>
attendance: about 55 people

(beer courtesy of NTT DoCoMo)

=============================

Based on interest from folks engaged in manet, mobileip, and seamoby
working groups, an unofficial meeting was held to discuss issues
surrounding the triggers at layer 2 to communicate information to
upper layers in a hosts protocol stack.  Two drafts, one discussing
the issues (draft-manyfolks-l2-mobilereq-01.txt) and another proposing
an API (draft-guri-seamoby-paging-trigger-00.txt) had been previously
submitted.  The objective of the meeting was to discover if the scope
should be more broad than the drafts proposed.  This was accomplished
by a somewhat unstructured discussion that touched on several topics
including what information to express, how it is delivered, and some
possible next steps.  The conversation and some perceived conclusions
are summarized below.

WHAT TO EXPRESS

Not a new problem, past experience likely useful:

The need for communication about L2 state to L3 is not new.  Routers have
been handling it, largely with proprietary means for years.  Routers are in
fact the canonical customers for this function; they by definition perform
their principal L3 function based on L2 state info.  Routing folks have
learned a lot about how links lie about their state (e.g., links that fail
half-duplex).

In fact, router manufacturers may have insight as to which metrics would be
of interest or possibly how to extend or abstract them if they are willing
to share the information.  Getting input from people familiar with routing
would be valuable.

Small set of common functions seen as useful:

The most common need expressed for information to be revealed was a
link up/down indicator.  This might be extended to be a table
containing reachability to multiple access points for technologies
where that is possible.  (There already exists an appropriate SNMP
message but it appears to be largely unused.)

Interaction with MobileIP:

In mobileIP on a wireless link, an L3 handover may, but does not always,
occur when the mobile node moves to a new access point.  Link layer
protocols provide more accurate, and possibly timelier, reachability
information than L3 'hello' protocols.  There is great benefit if a
mobileIP L3 can learn that a handover is imminent, rather than waiting
until after the old link is gone.  Early warning may allow mobileIP to
optimize the L3 hand-off process in a number of ways, such as triggering an
L3 early hand-off or moving context to the new router.

Discussion about architectural principles and amount of revealed information:

It was suggested that it might be valuable to reveal *any* information that
might be of interest to any part of the host, e.g., bit error rate, or link
bandwidth.  For example, a video codec (e.g., MPEG-4) might behave
differently if it was aware of the bit error rate of the wireless link.
This was widely thought to be a bad idea as it suggests an architecture
which is not general - leading to applications that are too tightly bound to
specific link types or depend on the assumption that the 'difficult' link
is the one nearest the end host.  The Internet layer, as the 'waist in the
protocol hourglass' insulates the application from the link layer.  The
question of what information ought to be passed up from L3 is separate from
what info L3 could usefully get from L2, and should be considered
separately.  Only a very limited, general set of parameters would be
appropriate above L3 and defining appropriate parameters based on link
characteristics is a very difficult task.  Even adding a metric such as
signal strength is difficult when link technologies are heterogeneous.

Some other caveats that were suggested: Triggers should only enable
performance enhancements and not be used for correct protocol
operation.  There's value in finding implicit rather than explicit
mechanisms.  Again: the focus should be firmly on L2/L3 communication,
rather than enabling applications (or, presumably, transport) to
interact directly with L2.  Apps can fight with L3 if it's done wrong.

Management applications considered a special case:

In the context of the above discussion, management applications were seen
as a special case that might require greater access to link information
than would normally be made available to normal applications (or even
protocol stack internals).  This is because the application is
intentionally associated with the management of a specific L2 technology.
For example, 802.11 cards typically come with a management application that
reveals information about the card status, link status, and access
point(s).  This information is necessary to ensure the correct functioning
of the card and should not be considered an Internet network application.

HOW IS L2 INFORMATION DELIVERED

Options for how to implement L2/3 communication include an API, a MIB, or a
protocol.  Most of the applications discussed above do not require
communication across the wire and so a protocol does not seem to be
justified or appropriate.  It's unclear whether an API or MIB is more
appropriate.  MIBs were considered by some as somewhat 'heavyweight'
because accessing local information typically requires SNMP to localhost
and implies ASN.1 marshaling that is not needed in this situation.  This
issue is particularly important to small memory and power-limited wireless
devices.

Most participants agreed that the difficult task that needs to be done is to
determine the right set of metrics to abstract and capture them in a common
location.  It was widely recognized that the question of MIB or API is
really secondary to the question of what information should be passed
between L2 and L3.  In the IETF, that process is frequently captured in a
MIB. However, since the issue is really a matter of internal communication
within a host, it may not be necessary to standardize or even specify the
format or mechanism which is used to communicate across the L2/L3 boundary.

SOME NEXT STEPS

Since the IETF doesn't develop link technologies, one output of an IETF
effort might include an informational RFC on which information to
export/reveal.  This document will likely be of interest to other standards
bodies who standardize link technologies.  (Some of these ideas have
already been taken to IEEE 802.11.  The IEEE is also working on related
issues in the context of OAM.) This might be considered a natural extension
of the PILC working group document "Advice to Subnet Designers."

The IESG will provide advice as to the next process steps.  They could
include doing the work in PILC, possibly requiring rechartering,
holding a BoF, creating a new working group, or possibly just an
individual RFC submission.  For the moment, the discussion should take
place on the PILC mail list.  Subscription instructions are available
through the working group charter page at
http://www.ietf.org/html.charters/pilc-charter.html.  Should the work
take place in a seperate working group, naturally a new mailing list
will be formed.

--aaron (for aaron and carl)

Postscript

In the wireless directorate meeting (of chairs of wireless-related working
groups, it was suggested that besides advising link designers, we should
also consider an advisory document to upper layers on how to handle L2
information.




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Apr  4 07:27:53 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA11399
	for <tsvwg-archive@odin.ietf.org>; Thu, 4 Apr 2002 07:27:53 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA25677
	for tsvwg-archive@odin.ietf.org; Thu, 4 Apr 2002 07:27:55 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA24803;
	Thu, 4 Apr 2002 07:04:56 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA24774
	for <tsvwg@optimus.ietf.org>; Thu, 4 Apr 2002 07:04:54 -0500 (EST)
Received: from irmgard.exp-math.uni-essen.de (ms.exp-math.uni-essen.de [132.252.150.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA10884
	for <tsvwg@ietf.org>; Thu, 4 Apr 2002 07:04:50 -0500 (EST)
Received: from there (biene.exp-math.uni-essen.de [132.252.150.215])
	by irmgard.exp-math.uni-essen.de (8.12.2/8.12.2) with SMTP id g34C4jsj022882;
	Thu, 4 Apr 2002 14:04:46 +0200
Message-Id: <200204041204.g34C4jsj022882@irmgard.exp-math.uni-essen.de>
Content-Type: text/plain;
  charset="iso-8859-1"
From: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
Organization: University Essen
To: Randall Stewart at work <rrs@cisco.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt
Date: Thu, 4 Apr 2002 14:04:42 +0200
X-Mailer: KMail [version 1.3.1]
References: <200203300410.g2U4AmC23317@newdev.harvard.edu> <3CA7A29E.B0676A86@cisco.com> <3CA8951B.DC94746F@cisco.com>
In-Reply-To: <3CA8951B.DC94746F@cisco.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Randy, dear all,

> The other choice (that is at least acceptable to me) is to leave
> the text alone :>
>
> What is the opinion of the WG??

I can just speak for myself, but I am perfectly happy with
the text as it is -- and from a practical point of view, there
is a sufficient number of open source implementations and monitoring
tools out there that may be used to test interop with...

This discussion has been sickening for a while, already.... 
get an RFC out !

Best regards,
Andreas
- -- 
- ------------------------------------------------------------------------
Dipl.-Ing. Andreas Jungmaier              |                         _
Computer Networking Technology Group      | ASCII ribbon campaign  ( )
University of Essen                       |  - against HTML email   X
http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8rEFaXAsLBNOCSsARAqRrAKCT1HCxBZSnJ2zrQ2wax/Fc97MSlACg9PBJ
qXlRWGvSofJoYAVGvzSkwK4=
=IZ/T
-----END PGP SIGNATURE-----

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Apr  4 14:35:50 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA05858
	for <tsvwg-archive@odin.ietf.org>; Thu, 4 Apr 2002 14:35:50 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA27026
	for tsvwg-archive@odin.ietf.org; Thu, 4 Apr 2002 14:35:50 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA25533;
	Thu, 4 Apr 2002 14:13:52 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA25506
	for <tsvwg@optimus.ietf.org>; Thu, 4 Apr 2002 14:13:50 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-193.dallas.net [209.44.42.193])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA04687
	for <tsvwg@ietf.org>; Thu, 4 Apr 2002 14:13:45 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id NAA16170;
	Thu, 4 Apr 2002 13:13:33 -0600
Date: Thu, 4 Apr 2002 13:13:32 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
Cc: Randall Stewart at work <rrs@cisco.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt
Message-ID: <20020404131332.D29566@openss7.org>
Reply-To: bidulock@openss7.org
References: <200203300410.g2U4AmC23317@newdev.harvard.edu> <3CA7A29E.B0676A86@cisco.com> <3CA8951B.DC94746F@cisco.com> <200204041204.g34C4jsj022882@irmgard.exp-math.uni-essen.de>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <200204041204.g34C4jsj022882@irmgard.exp-math.uni-essen.de>; from ajung@exp-math.uni-essen.de on Thu, Apr 04, 2002 at 02:04:42PM +0200
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Andreas,

Is it our opinion that the current text permits an implementation
to perform the calculation in any mathematically equivalent fashion
or using any meants that acheives the same result?

If so, why are you against saying so in the draft?

--brian


On Thu, 04 Apr 2002, Andreas Jungmaier wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Randy, dear all,
> 
> > The other choice (that is at least acceptable to me) is to leave
> > the text alone :>
> >
> > What is the opinion of the WG??
> 
> I can just speak for myself, but I am perfectly happy with
> the text as it is -- and from a practical point of view, there
> is a sufficient number of open source implementations and monitoring
> tools out there that may be used to test interop with...
> 
> This discussion has been sickening for a while, already.... 
> get an RFC out !
> 
> Best regards,
> Andreas
> - -- 
> - ------------------------------------------------------------------------
> Dipl.-Ing. Andreas Jungmaier              |                         _
> Computer Networking Technology Group      | ASCII ribbon campaign  ( )
> University of Essen                       |  - against HTML email   X
> http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.0.6 (GNU/Linux)
> Comment: For info see http://www.gnupg.org
> 
> iD8DBQE8rEFaXAsLBNOCSsARAqRrAKCT1HCxBZSnJ2zrQ2wax/Fc97MSlACg9PBJ
> qXlRWGvSofJoYAVGvzSkwK4=
> =IZ/T
> -----END PGP SIGNATURE-----
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Apr  4 15:26:08 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA07493
	for <tsvwg-archive@odin.ietf.org>; Thu, 4 Apr 2002 15:26:08 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA29667
	for tsvwg-archive@odin.ietf.org; Thu, 4 Apr 2002 15:26:10 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA27913;
	Thu, 4 Apr 2002 14:58:20 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA24586
	for <tsvwg@optimus.ietf.org>; Thu, 4 Apr 2002 07:02:31 -0500 (EST)
Received: from CNRI.Reston.VA.US (localhost [127.0.0.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA10755;
	Thu, 4 Apr 2002 07:02:28 -0500 (EST)
Message-Id: <200204041202.HAA10755@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: IETF-Announce: ;
Cc: tsvwg@ietf.org
From: Internet-Drafts@ietf.org
Reply-to: Internet-Drafts@ietf.org
Date: Thu, 04 Apr 2002 07:02:28 -0500
Subject: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-05.txt
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the Transport Area Working Group Working Group of the IETF.

	Title		: Stream Control Transmission Protocol (SCTP) Checksum 
                          Change
	Author(s)	: R. Stewart, J. Stone, D. Otis
	Filename	: draft-ietf-tsvwg-sctpcsum-05.txt
	Pages		: 14
	Date		: 03-Apr-02
	
SCTP [RFC2960] currently uses an Adler-32 checksum.  For small packets 
Adler-32 provides weak detection of errors.  This document changes that 
checksum and updates SCTP to use a 32 bit CRC checksum.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpcsum-05.txt

To remove yourself from the IETF Announcement list, send a message to 
ietf-announce-request with the word unsubscribe in the body of the message.

Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-tsvwg-sctpcsum-05.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-tsvwg-sctpcsum-05.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<20020403130524.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-tsvwg-sctpcsum-05.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-tsvwg-sctpcsum-05.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<20020403130524.I-D@ietf.org>

--OtherAccess--

--NextPart--




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Apr  5 04:57:34 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA00978
	for <tsvwg-archive@odin.ietf.org>; Fri, 5 Apr 2002 04:57:34 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id EAA19987
	for tsvwg-archive@odin.ietf.org; Fri, 5 Apr 2002 04:57:38 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA19395;
	Fri, 5 Apr 2002 04:40:04 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA19366
	for <tsvwg@optimus.ietf.org>; Fri, 5 Apr 2002 04:39:58 -0500 (EST)
Received: from irmgard.exp-math.uni-essen.de (irmgard.exp-math.uni-essen.de [132.252.150.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA00767
	for <tsvwg@ietf.org>; Fri, 5 Apr 2002 04:39:53 -0500 (EST)
Received: from biene.exp-math.uni-essen.de (biene.exp-math.uni-essen.de [132.252.150.215])
	by irmgard.exp-math.uni-essen.de (8.12.2/8.12.2) with ESMTP id g359dqsj026958;
	Fri, 5 Apr 2002 11:39:52 +0200
Content-Type: text/plain;
  charset="iso-8859-1"
From: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
Organization: University Essen
To: bidulock@openss7.org
Subject: Re: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt
Date: Fri, 5 Apr 2002 11:39:41 +0200
X-Mailer: KMail [version 1.4]
Cc: Randall Stewart at work <rrs@cisco.com>, tsvwg@ietf.org
References: <200203300410.g2U4AmC23317@newdev.harvard.edu> <200204041204.g34C4jsj022882@irmgard.exp-math.uni-essen.de> <20020404131332.D29566@openss7.org>
In-Reply-To: <20020404131332.D29566@openss7.org>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Message-Id: <200204051139.49793.ajung@exp-math.uni-essen.de>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thursday 04 April 2002 21:13, Brian F. G. Bidulock wrote:
> Andreas,
>
> Is it our opinion that the current text permits an implementation
> to perform the calculation in any mathematically equivalent fashion
> or using any meants that acheives the same result?
>
> If so, why are you against saying so in the draft?

Oops, 
I believe the busy authors of this draft have already made this
discussion about draft-ietf-tsvwg-sctpcsum-04.txt a thing of the
past....time to read the new draft !

Andreas

- ------------------------------------------------------------------------
Dipl.-Ing. Andreas Jungmaier              |                         _
Computer Networking Technology Group      | ASCII ribbon campaign  ( )
University of Essen                       |  - against HTML email   X
http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8rXDjXAsLBNOCSsARAmiqAKDqrxLY3sAYLZoU2Abu/cNBZlQliACg0C/l
u2WPXiXF5TLRTboRY3RCjRc=
=23sG
-----END PGP SIGNATURE-----


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Apr  5 07:35:45 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA02893
	for <tsvwg-archive@odin.ietf.org>; Fri, 5 Apr 2002 07:35:40 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA27433
	for tsvwg-archive@odin.ietf.org; Fri, 5 Apr 2002 07:35:43 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA26251;
	Fri, 5 Apr 2002 07:14:05 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA26227
	for <tsvwg@optimus.ietf.org>; Fri, 5 Apr 2002 07:14:02 -0500 (EST)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA02480
	for <tsvwg@ietf.org>; Fri, 5 Apr 2002 07:13:57 -0500 (EST)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g35CDwo84648
	for <tsvwg@ietf.org>; Fri, 5 Apr 2002 06:13:59 -0600 (CST)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CAD9506.ED5CC8FD@stewart.chicago.il.us>
Date: Fri, 05 Apr 2002 06:13:58 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: TSV WG list <tsvwg@ietf.org>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] Kernel Implementation NOW available
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi all:

Please note the FreeBSD/OpenBSD/NetBSD kernel reference implementation
version 1.0 is NOW available at:

http://www.sctp.org

This implementation as been tested at connect-a-thon and supports
the three BSD's mentioned above. Note for OpenBSD we have not
had a chance to test it :-0 Free and NetBSD have been tested at
connect-a-thon and the Minn IETF. 

Now what does this version support:

1) The full sockets API (both UDP and TCP models).
2) The PR-SCTP draft (formerly U-SCTP. PR-SCTP draft has not
                      been released .. soon.. but this
                      implementation fully supports the forthcoming
                      draft as agreed to at connect-a-thon and the
                      Minn IETF by interested folks at the bar-bof :>)
3) The latest sctp-addip draft.
4) The latest scoping draft.
5) The full implementors guide (current draft)
6) The CACC draft on SCTP.
7) Of course the new SCTP C-SUM aka using CRC32c. Note
   there is a compile option to have it use Adler32.

This version has been tested quite extensively on FreeBSD and NetBSD. It
appears quite stable but is still a "beta" release.

It is released under BSD liscense...

This implementation will eventually role into KAME and you will
then no longer need to download it from my sight once that
happens... For now you DO need to download KAME as well.

The 4/1/02 snapshot available at

http://www.kame.net

as been used and validated with this release.

You are required to be on:

FreeBSD 4.5
or
NetBSD 1.5.2
or
OpenBSD 3.0

Detailed build instructions are in the README file... a quick survey is
to

1) Install and build kame normally.

2) Create a symbolic link from your home directory
   to kame (where you placed the kame tree):

  ln -s /the/path/to/kame /home/me/kame

  [note if kame IS in your home directory skip this step]

3) Untar the file. Normally you do
   a) cd /home/me
   b) mkdir sctpBsd
   c) cd sctpBsd
   d) tar -zxf /path/to/kern_rel_1_0.tar.gz
   e) ./prepare.sh /home/me/sctpBsd

4) The prepare step will make all the needed symbolic links. Now
   go back into your GENERIC file that you used to build the
   kame tree and add "options SCTP". Note you MUST have BOTH
   INET and INET6 build as well. Do the syconfig and then the
   build again.... You now have SCTP in place. There are other
   steps you must do to get the libc updated... follow the
   instructions in the README or those generated by prepare.sh.



Please download and enjoy :>

(Any bugs or problems please send a email to 
 peterlei@cisco.com or rrs@cisco.com)


R
-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Apr  5 21:06:25 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA17503
	for <tsvwg-archive@odin.ietf.org>; Fri, 5 Apr 2002 21:06:25 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id VAA12771
	for tsvwg-archive@odin.ietf.org; Fri, 5 Apr 2002 21:06:31 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA11598;
	Fri, 5 Apr 2002 20:38:42 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA00991
	for <tsvwg@optimus.ietf.org>; Fri, 5 Apr 2002 08:58:10 -0500 (EST)
Received: from smtp.comcast.net (smtp.comcast.net [24.153.64.2])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA04890
	for <tsvwg@ietf.org>; Fri, 5 Apr 2002 08:58:05 -0500 (EST)
Received: from livingroom.acm.org
 (pcp01480838pcs.ncstl01.de.comcast.net [68.82.70.12])
 by mtaout04.icomcast.net (iPlanet Messaging Server 5.1 (built Feb  6 2002))
 with ESMTP id <0GU3006N6LG2T8@mtaout04.icomcast.net> for tsvwg@ietf.org; Fri,
 05 Apr 2002 08:57:39 -0500 (EST)
Date: Fri, 05 Apr 2002 09:07:04 -0500
From: Phillip Conrad <conrad@acm.org>
Subject: Re: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-sctpcsum-04.txt
In-reply-to: <200204051139.49793.ajung@exp-math.uni-essen.de>
X-Sender: phillipconrad@mail.comcast.net
To: tsvwg@ietf.org
Message-id: <5.1.0.14.2.20020405085957.04e1e720@joda.cis.temple.edu>
MIME-version: 1.0
X-Mailer: QUALCOMM Windows Eudora Version 5.1
Content-type: text/plain; charset=us-ascii; format=flowed
Content-transfer-encoding: 7BIT
References: <20020404131332.D29566@openss7.org>
 <200203300410.g2U4AmC23317@newdev.harvard.edu>
 <200204041204.g34C4jsj022882@irmgard.exp-math.uni-essen.de>
 <20020404131332.D29566@openss7.org>
Content-Transfer-Encoding: 7BIT
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7BIT

At 11:39 AM 4/5/2002 +0200, Andreas Jungmaier wrote:
>Oops,
>I believe the busy authors of this draft have already made this
>discussion about draft-ietf-tsvwg-sctpcsum-04.txt a thing of the
>past....time to read the new draft !
>
>Andreas

I did a diff on 04 vs. 05.   it appears that the only changes in the 05 
version are:
    (1) spelling out SCTP in the title
    (2) typo fix:  s/bit 0 of bit 0/bit 0 of byte 3/
    (3) typo fix: cut and paste error that duplicated some text has been fixed.

So, it is essentially the same draft in terms of content.

I do echo Andreas' earlier sentiment: we need to bring this matter to a 
successful conclusion soon.

Phill




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Apr  8 03:40:04 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA17060
	for <tsvwg-archive@odin.ietf.org>; Mon, 8 Apr 2002 03:40:04 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id DAA16177
	for tsvwg-archive@odin.ietf.org; Mon, 8 Apr 2002 03:40:06 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA15225;
	Mon, 8 Apr 2002 03:26:19 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA15194
	for <tsvwg@optimus.ietf.org>; Mon, 8 Apr 2002 03:26:17 -0400 (EDT)
Received: from caduceus.fm.intel.com (fmr02.intel.com [192.55.52.25])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA16871
	for <tsvwg@ietf.org>; Mon, 8 Apr 2002 03:26:13 -0400 (EDT)
Received: from petasus.fm.intel.com (petasus.fm.intel.com [10.1.192.37])
	by caduceus.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.37 2002/04/01 17:50:55 root Exp $) with ESMTP id g387PhH23335
	for <tsvwg@ietf.org>; Mon, 8 Apr 2002 07:25:43 GMT
Received: from fmsmsxvs041.fm.intel.com (fmsmsxv041-1.fm.intel.com [132.233.48.109])
	by petasus.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.15 2002/04/01 17:51:48 root Exp $) with SMTP id g387OqF20635
	for <tsvwg@ietf.org>; Mon, 8 Apr 2002 07:24:52 GMT
Received: from fmsmsx019.fm.intel.com ([132.233.42.130])
 by fmsmsxvs041.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002040800264410531
 for <tsvwg@ietf.org>; Mon, 08 Apr 2002 00:26:44 -0700
Received: by fmsmsx019.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <2LPAQS23>; Mon, 8 Apr 2002 00:26:15 -0700
Message-ID: <A1171C2850D8D4119D7B0002A508C76401698374@bgsmsx91.iind.intel.com>
From: "Khan, Mohammad" <mo_khan@trillium.com>
To: "'tsvwg@ietf.org'" <tsvwg@ietf.org>
Cc: "Pillai, Hari" <h_pillai@trillium.com>
Date: Mon, 8 Apr 2002 00:24:02 -0700 
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Subject: [Tsvwg] Retransmission Query
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi All,
Following is the querry regarding retransmission improvement described in
section 2.8 of
http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpguide-03.txt .
What is the difference between normal retransmission and fast
retransmission? 
In section 2.8.2 of the same draft where new section has been added for RFC
2960 section 7.2.4
	In step 3 it says that Ignore the value of cwnd while performing the
fast retransmission.
	In step 5 it says that Mark the data chunks as being fast
retransmitted and thus ineligible for a subsequentfast 	retransmit but only
those allowed by cwnd would be retransmitted.
	Do the above two statements contradict?

Regards
kashif khan

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Apr  8 04:25:21 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA17734
	for <tsvwg-archive@odin.ietf.org>; Mon, 8 Apr 2002 04:25:20 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id EAA18556
	for tsvwg-archive@odin.ietf.org; Mon, 8 Apr 2002 04:25:23 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA17706;
	Mon, 8 Apr 2002 04:11:10 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA17678
	for <tsvwg@optimus.ietf.org>; Mon, 8 Apr 2002 04:11:08 -0400 (EDT)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA17454
	for <tsvwg@ietf.org>; Mon, 8 Apr 2002 04:11:03 -0400 (EDT)
Received: from netlab.hcltech.com (arun.hclt-guindy.co.in [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id NAA21754;
	Mon, 8 Apr 2002 13:39:30 +0530
Message-ID: <3CB150A5.4F3CAD13@netlab.hcltech.com>
Date: Mon, 08 Apr 2002 13:41:17 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: "Khan, Mohammad" <mo_khan@trillium.com>
CC: "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>,
        Anusuya B Poonja <anu@netlab.hcltech.com>
Subject: Re: [Tsvwg] Retransmission Query
References: <A1171C2850D8D4119D7B0002A508C76401698374@bgsmsx91.iind.intel.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

hi kashif,
    Comments inline...

"Khan, Mohammad" wrote:

> Hi All,
> Following is the querry regarding retransmission improvement described in
> section 2.8 of
> http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpguide-03.txt .
> What is the difference between normal retransmission and fast
> retransmission?
> In section 2.8.2 of the same draft where new section has been added for RFC
> 2960 section 7.2.4
>         In step 3 it says that Ignore the value of cwnd while performing the
> fast retransmission.
>         In step 5 it says that Mark the data chunks as being fast
> retransmitted and thus ineligible for a subsequentfast  retransmit but only
> those allowed by cwnd would be retransmitted.
>         Do the above two statements contradict?

        I think the answer for the above question  could be understood once we
know the difference between "Packets marked for Fast Retransmission" and
"First Packet which is fast Retransmitted"......
    In step 3,  draft says about the "First Fast Retransmitted packet" and in
        step 5,  draft says about the " Waiting Marked Fast Retransmit packets"....

ie., Only the first packet sent  during Fast Retransmition should have the privilage
      to override the Congestion rules...... The rest of the data chunks which are
marked for Fast Retransmition , but couldnot be sent due to PMTU restriction,
will be waiting till the CWND allows it to send data....

I think there is no contradiction.....   Hope I didnt miss any point...


Thanks
-arun




>
>
> Regards
> kashif khan
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr  8 10:08:32 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA24245
	for <tsvwg-archive@odin.ietf.org>; Mon, 8 Apr 2002 10:08:27 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA07593
	for tsvwg-archive@odin.ietf.org; Mon, 8 Apr 2002 10:08:29 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA03881;
	Mon, 8 Apr 2002 09:23:25 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA03832
	for <tsvwg@ns.ietf.org>; Mon, 8 Apr 2002 09:23:21 -0400 (EDT)
Received: from mgw-x1.nokia.com (mgw-x1.nokia.com [131.228.20.21])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA22828
	for <tsvwg@ietf.org>; Mon, 8 Apr 2002 09:23:18 -0400 (EDT)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir05nok.ntc.nokia.com (esvir05nokt.ntc.nokia.com [172.21.143.37])
	by mgw-x1.nokia.com (Switch-2.2.1/Switch-2.2.0) with ESMTP id g38DMTu01403
	for <tsvwg@ietf.org>; Mon, 8 Apr 2002 16:22:29 +0300 (EET DST)
Received: from esebh001.NOE.Nokia.com (unverified) by esvir05nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5a22ce3d8fac158f25078@esvir05nok.ntc.nokia.com>;
 Mon, 8 Apr 2002 16:23:18 +0300
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh001.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Mon, 8 Apr 2002 16:23:17 +0300
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] Retransmission Query
Date: Mon, 8 Apr 2002 16:23:16 +0300
Message-ID: <58AEDE27D48F884285007ABF827FF26308EF60@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] Retransmission Query
Thread-Index: AcHe1QbH7yMrlomCS22xpoyBa3J5sAAKEYPg
To: <arun@netlab.hcltech.com>, <mo_khan@trillium.com>
Cc: <tsvwg@ietf.org>, <h_pillai@trillium.com>, <anu@netlab.hcltech.com>
X-OriginalArrivalTime: 08 Apr 2002 13:23:17.0364 (UTC) FILETIME=[8B545B40:01C1DF00]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id JAA03833
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi Arun, Kashif and the rest of the people of the list!

	Let's see if I can make things clearer...

	By the way, there is a 04 version since the end of January...

> hi kashif,
>     Comments inline...
> 
> "Khan, Mohammad" wrote:
> 
> > Hi All,
> > Following is the querry regarding retransmission 
> improvement described in
> > section 2.8 of
> > 
> http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpgu
> ide-03.txt .
> > What is the difference between normal retransmission and fast
> > retransmission?

	Well, there are two reasons why you should retransmit a TSN. One of them is due to the Fast Retransmit algorithm, and the other is due to T3 expiration... This second reason is the "normal" retransmission.

> > In section 2.8.2 of the same draft where new section has 
> been added for RFC
> > 2960 section 7.2.4
> >         In step 3 it says that Ignore the value of cwnd 
> while performing the
> > fast retransmission.

	The point is that when you apply the Fast Retransmission algorithm, you have to shrink your cwnd, so, normally, you won't have any space to send any TSN since your flight size will normally be bigger than your cwnd. But what you are trying to do is avoiding a T3 timer expiration, so in this case you don't care about the value of cwnd and you fast retransmit the TSN. BUT, in any case, you only send one MTU worth of data...

> >         In step 5 it says that Mark the data chunks as being fast
> > retransmitted and thus ineligible for a subsequentfast  
> retransmit but only
> > those allowed by cwnd would be retransmitted.
> >         Do the above two statements contradict?

	No...

	One TSN that has been marked for Fast retransmission once, cannot be set for fast retransmission any more. This is because in a normal flow of data, there are both DATA chunks and SACK chunks in flight. So, it is likely that right after fast retransmitting a TSN, you will still receive many SACKs triggered by TSNs sent before the one you fast retransmitted. These SACKs will still include the gap, and thus you would most probably fast retransmit a TSN several times before you receive the SACK acknowledging the receipt of the first fast retransmission (acknowledging the gap). So, this is why one TSN can only be mark for retransmission once due to the Fast Retransmit algorithm.

	And about the second statement. A SACK chunk can cause several TSNs to be fast retransmitted. However, you should only send one MTU worth of data. This causes that you might have marked several TSNs for retransmission due to the fast retransmit algorithm, but you only send, let's say, one of them. The rest are marked for retransmission, but cannot be sent in that very moment. So, they will be retransmitted later, when cwnd allows to do so...
	

>         I think the answer for the above question  could be 
> understood once we
> know the difference between "Packets marked for Fast 
> Retransmission" and
> "First Packet which is fast Retransmitted"......
>     In step 3,  draft says about the "First Fast 
> Retransmitted packet" and in
>         step 5,  draft says about the " Waiting Marked Fast 
> Retransmit packets"....
> 
> ie., Only the first packet sent  during Fast Retransmition 
> should have the privilage
>       to override the Congestion rules...... The rest of the 
> data chunks which are
> marked for Fast Retransmition , but couldnot be sent due to 
> PMTU restriction,
> will be waiting till the CWND allows it to send data....
> 
> I think there is no contradiction.....   Hope I didnt miss 
> any point...

	You are right. The point is that a TSN that has been marked for retransmission due to the fast retransmit algorithm is still a TSN that is marked for retransmission. It is like when the T3 expires: you might not be able to send all the TSNs marked for retransmission in the very moment when the T3 expires, however, they are kept marked for retransmission and will be resent later, when the cwnd allows to do so...

	I hope now things are a little bit clearer...

	BR Iván Arias Rodríguez

> Thanks
> -arun
> 
> 
> 
> 
> >
> >
> > Regards
> > kashif khan
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> --
> ****************************************************************
> V.Arun Prasad
> HCL Technologies Ltd.
> 51, Jawaharlal Nehru Road,
> Ekkattuthangal,
> Guindy Industrial Estate,
> Chennai - 600097.
> 
> Contact # : 9144 - 2334174
>             9144 - 2334181
>             extn : 233
> ****************************************************************
> 
> 
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr  9 01:26:12 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA18294
	for <tsvwg-archive@odin.ietf.org>; Tue, 9 Apr 2002 01:26:12 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id BAA20165
	for tsvwg-archive@odin.ietf.org; Tue, 9 Apr 2002 01:26:14 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id AAA08487;
	Tue, 9 Apr 2002 00:51:39 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id AAA08459
	for <tsvwg@optimus.ietf.org>; Tue, 9 Apr 2002 00:51:37 -0400 (EDT)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA17877
	for <tsvwg@ietf.org>; Tue, 9 Apr 2002 00:51:27 -0400 (EDT)
Received: from netlab.hcltech.com (arun.hclt-guindy.co.in [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id KAA08569;
	Tue, 9 Apr 2002 10:07:59 +0530
Message-ID: <3CB27092.7680BC16@netlab.hcltech.com>
Date: Tue, 09 Apr 2002 10:09:47 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: tsvwg <tsvwg@ietf.org>
CC: tcp-impl@grc.nasa.gov, sctp <sctpimp@netlab.hcltech.com>
References: <3CADF78E.D9E24D26@cisco.com> <3CB20DBE.3B687309@cisco.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] ECN & PMTU
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi all,
    Some doubts on the procedure followed for PMTU (Path mtu) discovery and
ECN method.....
    The doubt is not related to the procedure as such, but the relation between the
two procedures stated above......
Tcp and Sctp uses ICMP message to handle the PMTU discovery procedure....
ie., the intermediate router will send an ICMP messge of "Message too Long"
to the end point sending the data packet of size greater than the MTU size of some
intermediate path to the destination.....

For ECN, the intermediate router should set the appropriate bits in the IP header
(if it supports ECN ....) whenever there is congestion and risk of packet drops
is high....... The endpoints then correct their Congestion window accordingly and
help in reducing the Congestion in the intermediate router.....

Why, we adopt two different methods to carry out similar (in the sense, both are
related to the Intermediate routers....)?????
    Why cant we generalise this stuff.... ie.,  the need here is some way to
communicate between the intermediate router and the endpoint.... Not sure
which of the two ways is advantageous, but cant we use the same method for
both..... ie.,
As for the PMTU discovery, the intermediate router can send an ICMP message
to the endpoint carrying a message "CONGESTION", by receiving this the
endpoint will do all appropriate actions as it does when it receives the packet set
with ECN-ECHO flag (as in present ECN procedure...)

or vice versa ( ie., adopt the method followed by ECN  for PMTU discovery aslo,
that might be slighly tough, considering the Backward compatibility....)

    What we can achieve by this is the simplicity  in the router implementation,
which doesnt need to do different procedure for different features of the Transport
layer..... and by maintaining this uniformity, the future extentions which demands a
similar requirement can use the same way......

I could have missed some points....

Any comments on this????


Thanks
-arun
--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr  9 02:40:33 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA27565
	for <tsvwg-archive@odin.ietf.org>; Tue, 9 Apr 2002 02:40:33 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id CAA23722
	for tsvwg-archive@odin.ietf.org; Tue, 9 Apr 2002 02:40:35 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id CAA22387;
	Tue, 9 Apr 2002 02:11:11 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id CAA22355
	for <tsvwg@optimus.ietf.org>; Tue, 9 Apr 2002 02:11:08 -0400 (EDT)
Received: from auemail1.firewall.lucent.com (auemail1.lucent.com [192.11.223.161])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA27208
	for <tsvwg@ietf.org>; Tue, 9 Apr 2002 02:11:06 -0400 (EDT)
Received: from ihrh1.emsr.lucent.com (h135-1-218-53.lucent.com [135.1.218.53])
	by auemail1.firewall.lucent.com (Switch-2.1.3/Switch-2.1.0) with ESMTP id g396A0X18994;
	Tue, 9 Apr 2002 02:10:00 -0400 (EDT)
Received: from new-wopr.eng.ascend.com by ihrh1.emsr.lucent.com (8.8.8+Sun/EMS-1.5 Solaris/emsr)
	id BAA23394 for ; Tue, 9 Apr 2002 01:09:57 -0500 (CDT)
Received: from grigri.eng.ascend.com (grigri.eng.ascend.com [135.140.53.45])
	by new-wopr.eng.ascend.com (8.10.2+Sun/8.10.2) with ESMTP id g3969wP17631;
	Mon, 8 Apr 2002 23:09:58 -0700 (PDT)
Received: from lucent.com by grigri.eng.ascend.com (8.8.8+Sun/SMI-SVR4)
	id XAA00287; Mon, 8 Apr 2002 23:09:57 -0700 (PDT)
Message-ID: <3CB285B4.9E8FBBA4@lucent.com>
Date: Mon, 08 Apr 2002 23:09:56 -0700
From: Jacob Heitz <jheitz@lucent.com>
Organization: Lucent Technologies, Internetworking Systems, Software Engineering, 
 Alameda
X-Mailer: Mozilla 4.76 [en] (X11; U; SunOS 5.8 sun4u)
X-Accept-Language: en
MIME-Version: 1.0
To: Arun Prasad <arun@netlab.hcltech.com>
CC: tsvwg <tsvwg@ietf.org>, tcp-impl@grc.nasa.gov,
        sctp <sctpimp@netlab.hcltech.com>
Subject: Re: [Tsvwg] ECN & PMTU
References: <3CADF78E.D9E24D26@cisco.com> <3CB20DBE.3B687309@cisco.com> <3CB27092.7680BC16@netlab.hcltech.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

One reason: sending extra packets in the face
of congestion will worsen the congestion.
Congestion collapse could be a very serious possible
consequence of sending extra packets.
Notice the ECN does not go towards the sender,
but in the other direction. I think the intention
is to have the upper layer (transport) in the
receiver withhold acknowledgements to cause the
sender to back off. I'm making this up, so please
correct me if I'm wrong.

PMTU discovery does not face this problem. It's ok to send
an extra packet back to the sender in this case.

Arun Prasad wrote:
> 
> Hi all,
>     Some doubts on the procedure followed for PMTU (Path mtu) discovery and
> ECN method.....
>     The doubt is not related to the procedure as such, but the relation between the
> two procedures stated above......
> Tcp and Sctp uses ICMP message to handle the PMTU discovery procedure....
> ie., the intermediate router will send an ICMP messge of "Message too Long"
> to the end point sending the data packet of size greater than the MTU size of some
> intermediate path to the destination.....
> 
> For ECN, the intermediate router should set the appropriate bits in the IP header
> (if it supports ECN ....) whenever there is congestion and risk of packet drops
> is high....... The endpoints then correct their Congestion window accordingly and
> help in reducing the Congestion in the intermediate router.....
> 
> Why, we adopt two different methods to carry out similar (in the sense, both are
> related to the Intermediate routers....)?????
>     Why cant we generalise this stuff.... ie.,  the need here is some way to
> communicate between the intermediate router and the endpoint.... Not sure
> which of the two ways is advantageous, but cant we use the same method for
> both..... ie.,
> As for the PMTU discovery, the intermediate router can send an ICMP message
> to the endpoint carrying a message "CONGESTION", by receiving this the
> endpoint will do all appropriate actions as it does when it receives the packet set
> with ECN-ECHO flag (as in present ECN procedure...)
> 
> or vice versa ( ie., adopt the method followed by ECN  for PMTU discovery aslo,
> that might be slighly tough, considering the Backward compatibility....)
> 
>     What we can achieve by this is the simplicity  in the router implementation,
> which doesnt need to do different procedure for different features of the Transport
> layer..... and by maintaining this uniformity, the future extentions which demands a
> similar requirement can use the same way......
> 
> I could have missed some points....
> 
> Any comments on this????
> 
> Thanks
> -arun
> --
> ****************************************************************
> V.Arun Prasad
> HCL Technologies Ltd.
> 51, Jawaharlal Nehru Road,
> Ekkattuthangal,
> Guindy Industrial Estate,
> Chennai - 600097.
> 
> Contact # : 9144 - 2334174
>             9144 - 2334181
>             extn : 233
> ****************************************************************
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
// Jacob Heitz. Tel:510-747-2917

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr  9 05:23:43 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA00189
	for <tsvwg-archive@odin.ietf.org>; Tue, 9 Apr 2002 05:23:43 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA03604
	for tsvwg-archive@odin.ietf.org; Tue, 9 Apr 2002 05:23:46 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA02748;
	Tue, 9 Apr 2002 05:02:48 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA02715
	for <tsvwg@optimus.ietf.org>; Tue, 9 Apr 2002 05:02:45 -0400 (EDT)
Received: from baqaqi.chi.il.us (12-249-2-251.client.attbi.com [12.249.2.251])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA29903
	for <tsvwg@ietf.org>; Tue, 9 Apr 2002 05:02:42 -0400 (EDT)
Received: from baqaqi.chi.il.us (piggy@localhost)
	by baqaqi.chi.il.us (8.11.0/8.11.0) with ESMTP id g3991Q705601;
	Tue, 9 Apr 2002 04:01:31 -0500
Message-Id: <200204090901.g3991Q705601@baqaqi.chi.il.us>
To: Jacob Heitz <jheitz@lucent.com>
cc: Arun Prasad <arun@netlab.hcltech.com>, tsvwg <tsvwg@ietf.org>,
        tcp-impl@grc.nasa.gov, sctp <sctpimp@netlab.hcltech.com>
From: "La Monte Henry Piggy Yarroll" <piggy@baqaqi.chi.il.us>
Subject: Re: [Tsvwg] ECN & PMTU 
In-reply-to: Your message of Mon, 08 Apr 2002 23:09:56 -0700.
             <3CB285B4.9E8FBBA4@lucent.com> 
Date: Tue, 09 Apr 2002 04:01:26 -0500
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Jacob's comments are roughly correct.  Bits in the IP header are very
precious so we expend them only for a REALLY good reason, e.g. ECN.

Jacob Heitz <jheitz@lucent.com> writes:
> One reason: sending extra packets in the face
> of congestion will worsen the congestion.
> Congestion collapse could be a very serious possible
> consequence of sending extra packets.
> Notice the ECN does not go towards the sender,
> but in the other direction. I think the intention
> is to have the upper layer (transport) in the
> receiver withhold acknowledgements to cause the
> sender to back off. I'm making this up, so please
> correct me if I'm wrong.

You are correct that the transport layer is supposed to handle the
Congestion Experienced (CE) bits in the IP header.  In the case of
both SCTP and TCP the reaction is to include a transport-specific
congestion notification in the next ACK.  This allows the sending side
to back off just like normal congestion detection, but without losing
a packet.

> PMTU discovery does not face this problem. It's ok to send
> an extra packet back to the sender in this case.

I.e. there is no need to spend IP header bits on this function.

> Arun Prasad wrote:
> > 
> > Hi all,
> >     Some doubts on the procedure followed for PMTU (Path mtu) discovery and
> > ECN method.....
> >     The doubt is not related to the procedure as such, but the relation bet
ween the
> > two procedures stated above......
> > Tcp and Sctp uses ICMP message to handle the PMTU discovery procedure....
> > ie., the intermediate router will send an ICMP messge of "Message too Long"
> > to the end point sending the data packet of size greater than the MTU size 
of some
> > intermediate path to the destination.....
> > 
> > For ECN, the intermediate router should set the appropriate bits in the IP 
header
> > (if it supports ECN ....) whenever there is congestion and risk of packet d
rops
> > is high....... The endpoints then correct their Congestion window according
ly and
> > help in reducing the Congestion in the intermediate router.....
> > 
> > Why, we adopt two different methods to carry out similar (in the sense, bot
h are
> > related to the Intermediate routers....)?????
> >     Why cant we generalise this stuff.... ie.,  the need here is some way t
o
> > communicate between the intermediate router and the endpoint.... Not sure
> > which of the two ways is advantageous, but cant we use the same method for
> > both..... ie.,
> > As for the PMTU discovery, the intermediate router can send an ICMP message
> > to the endpoint carrying a message "CONGESTION", by receiving this the
> > endpoint will do all appropriate actions as it does when it receives the pa
cket set
> > with ECN-ECHO flag (as in present ECN procedure...)
> > 
> > or vice versa ( ie., adopt the method followed by ECN  for PMTU discovery a
slo,
> > that might be slighly tough, considering the Backward compatibility....)
> > 
> >     What we can achieve by this is the simplicity  in the router implementa
tion,
> > which doesnt need to do different procedure for different features of the T
ransport
> > layer..... and by maintaining this uniformity, the future extentions which 
demands a
> > similar requirement can use the same way......
> > 
> > I could have missed some points....
> > 
> > Any comments on this????
> > 
> > Thanks
> > -arun
> > --
> > ****************************************************************
> > V.Arun Prasad
> > HCL Technologies Ltd.
> > 51, Jawaharlal Nehru Road,
> > Ekkattuthangal,
> > Guindy Industrial Estate,
> > Chennai - 600097.
> > 
> > Contact # : 9144 - 2334174
> >             9144 - 2334181
> >             extn : 233
> > ****************************************************************
> > 
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> -- 
> // Jacob Heitz. Tel:510-747-2917
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr  9 06:13:57 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA00992
	for <tsvwg-archive@odin.ietf.org>; Tue, 9 Apr 2002 06:13:57 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA06931
	for tsvwg-archive@odin.ietf.org; Tue, 9 Apr 2002 06:14:00 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA05684;
	Tue, 9 Apr 2002 05:57:15 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA05655
	for <tsvwg@optimus.ietf.org>; Tue, 9 Apr 2002 05:57:13 -0400 (EDT)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA00640
	for <tsvwg@ietf.org>; Tue, 9 Apr 2002 05:57:08 -0400 (EDT)
Received: from netlab.hcltech.com (arun.hclt-guindy.co.in [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id PAA15500;
	Tue, 9 Apr 2002 15:25:35 +0530
Message-ID: <3CB2BB02.21AA779C@netlab.hcltech.com>
Date: Tue, 09 Apr 2002 15:27:22 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Jacob Heitz <jheitz@lucent.com>
CC: tsvwg <tsvwg@ietf.org>, tcp-impl@grc.nasa.gov,
        sctp <sctpimp@netlab.hcltech.com>
Subject: Re: [Tsvwg] ECN & PMTU
References: <3CADF78E.D9E24D26@cisco.com> <3CB20DBE.3B687309@cisco.com> <3CB27092.7680BC16@netlab.hcltech.com> <3CB285B4.9E8FBBA4@lucent.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

hi jacob,

Jacob Heitz wrote:

> One reason: sending extra packets in the face
> of congestion will worsen the congestion.
> Congestion collapse could be a very serious possible
> consequence of sending extra packets.

         I couldnt get the point of, sending extra packets during congestion.....
Say if a Router decides that there will be congestion ( its not in congestion now...)
then it decides to set the ECN bit in the IP header of certain packets in the Queue...
These packets should reach the receiver endpoint.... then the receiver will
send a ACK to the sender to indicate there is a congestion..... Doesnt this too
long a process.... ie., anyway a SACK is to be sent from the Receiver to the
sender....  In that case why not an ICMP packet from the router to the sender???

And also, since the extra packets are going to flow in the opposite direction
of the actual congestion, I couldnt see any harm .... (I know this might be a
wrong statemnent.... but I dont have the right justification...)

Atleast in Source Quenching method, the flooding of Data by the Sender can be
controlled fast than the current ECN method...right!!


>
> Notice the ECN does not go towards the sender,
> but in the other direction. I think the intention
> is to have the upper layer (transport) in the
> receiver withhold acknowledgements to cause the
> sender to back off. I'm making this up, so please

        Why at all a receiver bother about the Congetion....  I see no reasons for the
receiver to know about the congestion in the path.... because he cant do anything
about that...
And also by doing this by Source Quenching method.... its not mandatory that
both the source and destination Transport layers support the ECN capability..
If only sender is supporting, its more than enough... and there will be no necessity
to negotiate the support of this feature, end to end....
Router  also doesnt bother  whether an Endpoint is supporting  ECN or not...
he can just send an ICMP message to the sender, whether he supports ECN or not...
(slightly costly operation but thought its ok...)

Thanks
-arun

>
> correct me if I'm wrong.
>
> PMTU discovery does not face this problem. It's ok to send
> an extra packet back to the sender in this case.
>
> Arun Prasad wrote:
> >
> > Hi all,
> >     Some doubts on the procedure followed for PMTU (Path mtu) discovery and
> > ECN method.....
> >     The doubt is not related to the procedure as such, but the relation between the
> > two procedures stated above......
> > Tcp and Sctp uses ICMP message to handle the PMTU discovery procedure....
> > ie., the intermediate router will send an ICMP messge of "Message too Long"
> > to the end point sending the data packet of size greater than the MTU size of some
> > intermediate path to the destination.....
> >
> > For ECN, the intermediate router should set the appropriate bits in the IP header
> > (if it supports ECN ....) whenever there is congestion and risk of packet drops
> > is high....... The endpoints then correct their Congestion window accordingly and
> > help in reducing the Congestion in the intermediate router.....
> >
> > Why, we adopt two different methods to carry out similar (in the sense, both are
> > related to the Intermediate routers....)?????
> >     Why cant we generalise this stuff.... ie.,  the need here is some way to
> > communicate between the intermediate router and the endpoint.... Not sure
> > which of the two ways is advantageous, but cant we use the same method for
> > both..... ie.,
> > As for the PMTU discovery, the intermediate router can send an ICMP message
> > to the endpoint carrying a message "CONGESTION", by receiving this the
> > endpoint will do all appropriate actions as it does when it receives the packet set
> > with ECN-ECHO flag (as in present ECN procedure...)
> >
> > or vice versa ( ie., adopt the method followed by ECN  for PMTU discovery aslo,
> > that might be slighly tough, considering the Backward compatibility....)
> >
> >     What we can achieve by this is the simplicity  in the router implementation,
> > which doesnt need to do different procedure for different features of the Transport
> > layer..... and by maintaining this uniformity, the future extentions which demands a
> > similar requirement can use the same way......
> >
> > I could have missed some points....
> >
> > Any comments on this????
> >
> > Thanks
> > -arun
> > --
> > ****************************************************************
> > V.Arun Prasad
> > HCL Technologies Ltd.
> > 51, Jawaharlal Nehru Road,
> > Ekkattuthangal,
> > Guindy Industrial Estate,
> > Chennai - 600097.
> >
> > Contact # : 9144 - 2334174
> >             9144 - 2334181
> >             extn : 233
> > ****************************************************************
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
>
> --
> // Jacob Heitz. Tel:510-747-2917

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr  9 06:50:17 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA01510
	for <tsvwg-archive@odin.ietf.org>; Tue, 9 Apr 2002 06:50:17 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA08710
	for tsvwg-archive@odin.ietf.org; Tue, 9 Apr 2002 06:50:18 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA07558;
	Tue, 9 Apr 2002 06:27:16 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA07528
	for <tsvwg@optimus.ietf.org>; Tue, 9 Apr 2002 06:27:14 -0400 (EDT)
Received: from ihemail2.firewall.lucent.com (ihemail2.lucent.com [192.11.222.163])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA01135
	for <tsvwg@ietf.org>; Tue, 9 Apr 2002 06:27:12 -0400 (EDT)
Received: from ihrh1.emsr.lucent.com (h135-1-218-53.lucent.com [135.1.218.53])
	by ihemail2.firewall.lucent.com (Switch-2.1.3/Switch-2.1.0) with ESMTP id g39AR7h14262;
	Tue, 9 Apr 2002 06:27:07 -0400 (EDT)
Received: from new-wopr.eng.ascend.com by ihrh1.emsr.lucent.com (8.8.8+Sun/EMS-1.5 Solaris/emsr)
	id FAA13335 for ; Tue, 9 Apr 2002 05:27:05 -0500 (CDT)
Received: from grigri.eng.ascend.com (grigri.eng.ascend.com [135.140.53.45])
	by new-wopr.eng.ascend.com (8.10.2+Sun/8.10.2) with ESMTP id g39AR5P26790;
	Tue, 9 Apr 2002 03:27:05 -0700 (PDT)
Received: from lucent.com by grigri.eng.ascend.com (8.8.8+Sun/SMI-SVR4)
	id DAA11594; Tue, 9 Apr 2002 03:27:04 -0700 (PDT)
Message-ID: <3CB2C285.FA4806AB@lucent.com>
Date: Tue, 09 Apr 2002 03:29:25 -0700
From: Jacob Heitz <jheitz@lucent.com>
Organization: Lucent Technologies, Internetworking Systems, Software Engineering, 
 Alameda
X-Mailer: Mozilla 4.77 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Arun Prasad <arun@netlab.hcltech.com>
CC: tsvwg <tsvwg@ietf.org>, tcp-impl@grc.nasa.gov,
        sctp <sctpimp@netlab.hcltech.com>
Subject: Re: [Tsvwg] ECN & PMTU
References: <3CADF78E.D9E24D26@cisco.com> <3CB20DBE.3B687309@cisco.com> <3CB27092.7680BC16@netlab.hcltech.com> <3CB285B4.9E8FBBA4@lucent.com> <3CB2BB02.21AA779C@netlab.hcltech.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

First of all, I gave a reason why it is so.
If you think another way is better,
that's nice, but you won't change anything.

Source quenching was always controversial.
It is considered higher impact, because it adds
an extra packet to the network. Setting a CE
bit adds nothing that was not there before anyway,
so is considered to be safer, or lower impact.

Yes, it's slower than having the congested router
send an ICMP.

The end receiver sends an ACK anyway. Let it modify
or withhold that ACK. This is what the receiver can
do. It would send the ACK anyway, thus nothing
extra is added to the network.

But consider these points:

A network is made of different makes/models of routers.
A router may consider itself congested in all
directions at once when it runs low on buffer memory.
Then it may send ICMPs every which way.

If a path is congested at many routers, each may send
an ICMP in the reverse direction, resulting in a
multitude of ICMPs reaching the sender.

When congestion collapse looms, safety first is the
golden rule: add nothing more that was not already
there anyway.

Arun Prasad wrote:
> 
> hi jacob,
> 
> Jacob Heitz wrote:
> 
> > One reason: sending extra packets in the face
> > of congestion will worsen the congestion.
> > Congestion collapse could be a very serious possible
> > consequence of sending extra packets.
> 
>          I couldnt get the point of, sending extra packets during congestion.....
> Say if a Router decides that there will be congestion ( its not in congestion now...)
> then it decides to set the ECN bit in the IP header of certain packets in the Queue...
> These packets should reach the receiver endpoint.... then the receiver will
> send a ACK to the sender to indicate there is a congestion..... Doesnt this too
> long a process.... ie., anyway a SACK is to be sent from the Receiver to the
> sender....  In that case why not an ICMP packet from the router to the sender???
> 
> And also, since the extra packets are going to flow in the opposite direction
> of the actual congestion, I couldnt see any harm .... (I know this might be a
> wrong statemnent.... but I dont have the right justification...)
> 
> Atleast in Source Quenching method, the flooding of Data by the Sender can be
> controlled fast than the current ECN method...right!!
> 
> >
> > Notice the ECN does not go towards the sender,
> > but in the other direction. I think the intention
> > is to have the upper layer (transport) in the
> > receiver withhold acknowledgements to cause the
> > sender to back off. I'm making this up, so please
> 
>         Why at all a receiver bother about the Congetion....  I see no reasons for the
> receiver to know about the congestion in the path.... because he cant do anything
> about that...
> And also by doing this by Source Quenching method.... its not mandatory that
> both the source and destination Transport layers support the ECN capability..
> If only sender is supporting, its more than enough... and there will be no necessity
> to negotiate the support of this feature, end to end....
> Router  also doesnt bother  whether an Endpoint is supporting  ECN or not...
> he can just send an ICMP message to the sender, whether he supports ECN or not...
> (slightly costly operation but thought its ok...)
> 
> Thanks
> -arun
> 
> >
> > correct me if I'm wrong.
> >
> > PMTU discovery does not face this problem. It's ok to send
> > an extra packet back to the sender in this case.
> >
> > Arun Prasad wrote:
> > >
> > > Hi all,
> > >     Some doubts on the procedure followed for PMTU (Path mtu) discovery and
> > > ECN method.....
> > >     The doubt is not related to the procedure as such, but the relation between the
> > > two procedures stated above......
> > > Tcp and Sctp uses ICMP message to handle the PMTU discovery procedure....
> > > ie., the intermediate router will send an ICMP messge of "Message too Long"
> > > to the end point sending the data packet of size greater than the MTU size of some
> > > intermediate path to the destination.....
> > >
> > > For ECN, the intermediate router should set the appropriate bits in the IP header
> > > (if it supports ECN ....) whenever there is congestion and risk of packet drops
> > > is high....... The endpoints then correct their Congestion window accordingly and
> > > help in reducing the Congestion in the intermediate router.....
> > >
> > > Why, we adopt two different methods to carry out similar (in the sense, both are
> > > related to the Intermediate routers....)?????
> > >     Why cant we generalise this stuff.... ie.,  the need here is some way to
> > > communicate between the intermediate router and the endpoint.... Not sure
> > > which of the two ways is advantageous, but cant we use the same method for
> > > both..... ie.,
> > > As for the PMTU discovery, the intermediate router can send an ICMP message
> > > to the endpoint carrying a message "CONGESTION", by receiving this the
> > > endpoint will do all appropriate actions as it does when it receives the packet set
> > > with ECN-ECHO flag (as in present ECN procedure...)
> > >
> > > or vice versa ( ie., adopt the method followed by ECN  for PMTU discovery aslo,
> > > that might be slighly tough, considering the Backward compatibility....)
> > >
> > >     What we can achieve by this is the simplicity  in the router implementation,
> > > which doesnt need to do different procedure for different features of the Transport
> > > layer..... and by maintaining this uniformity, the future extentions which demands a
> > > similar requirement can use the same way......
> > >
> > > I could have missed some points....
> > >
> > > Any comments on this????
> > >
> > > Thanks
> > > -arun
> > > --
> > > ****************************************************************
> > > V.Arun Prasad
> > > HCL Technologies Ltd.
> > > 51, Jawaharlal Nehru Road,
> > > Ekkattuthangal,
> > > Guindy Industrial Estate,
> > > Chennai - 600097.
> > >
> > > Contact # : 9144 - 2334174
> > >             9144 - 2334181
> > >             extn : 233
> > > ****************************************************************
> > >
> > > _______________________________________________
> > > tsvwg mailing list
> > > tsvwg@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/tsvwg
> >
> > --
> > // Jacob Heitz. Tel:510-747-2917
> 
> --
> ****************************************************************
> V.Arun Prasad
> HCL Technologies Ltd.
> 51, Jawaharlal Nehru Road,
> Ekkattuthangal,
> Guindy Industrial Estate,
> Chennai - 600097.
> 
> Contact # : 9144 - 2334174
>             9144 - 2334181
>             extn : 233
> ****************************************************************

-- 
// Jacob Heitz. Tel:510-747-2917

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr  9 07:40:37 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA02280
	for <tsvwg-archive@odin.ietf.org>; Tue, 9 Apr 2002 07:40:37 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA12158
	for tsvwg-archive@odin.ietf.org; Tue, 9 Apr 2002 07:40:39 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA10626;
	Tue, 9 Apr 2002 07:18:25 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA10536
	for <tsvwg@optimus.ietf.org>; Tue, 9 Apr 2002 07:18:19 -0400 (EDT)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA01979
	for <tsvwg@ietf.org>; Tue, 9 Apr 2002 07:18:16 -0400 (EDT)
Received: from netlab.hcltech.com (arun.hclt-guindy.co.in [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id QAA16947;
	Tue, 9 Apr 2002 16:46:36 +0530
Message-ID: <3CB2CDFF.2F583DF1@netlab.hcltech.com>
Date: Tue, 09 Apr 2002 16:48:23 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Jacob Heitz <jheitz@lucent.com>
CC: tsvwg <tsvwg@ietf.org>, tcp-impl@grc.nasa.gov,
        sctp <sctpimp@netlab.hcltech.com>
Subject: Re: [Tsvwg] ECN & PMTU
References: <3CADF78E.D9E24D26@cisco.com> <3CB20DBE.3B687309@cisco.com> <3CB27092.7680BC16@netlab.hcltech.com> <3CB285B4.9E8FBBA4@lucent.com> <3CB2BB02.21AA779C@netlab.hcltech.com> <3CB2C285.FA4806AB@lucent.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

hi jacob,


Jacob Heitz wrote:

> First of all, I gave a reason why it is so.
> If you think another way is better,
> that's nice, but you won't change anything.
>

    thats slightly harsh.....

>
> Source quenching was always controversial.
> It is considered higher impact, because it adds
> an extra packet to the network. Setting a CE
> bit adds nothing that was not there before anyway,
> so is considered to be safer, or lower impact.
>
> Yes, it's slower than having the congested router
> send an ICMP.
>
> The end receiver sends an ACK anyway. Let it modify
> or withhold that ACK. This is what the receiver can
> do. It would send the ACK anyway, thus nothing
> extra is added to the network.
>
> But consider these points:
>
> A network is made of different makes/models of routers.
> A router may consider itself congested in all
> directions at once when it runs low on buffer memory.
> Then it may send ICMPs every which way.
>
> If a path is congested at many routers, each may send
> an ICMP in the reverse direction, resulting in a
> multitude of ICMPs reaching the sender.
>

    Ya, I got the point.... The above point really justifies the reason of not
selecting Source Quenching for ECN..... ( That may be obvious, but I missed
it....).. Thanks for this...

>
> When congestion collapse looms, safety first is the
> golden rule: add nothing more that was not already
> there anyway.
>


Thanks
-arun

>
> Arun Prasad wrote:
> >
> > hi jacob,
> >
> > Jacob Heitz wrote:
> >
> > > One reason: sending extra packets in the face
> > > of congestion will worsen the congestion.
> > > Congestion collapse could be a very serious possible
> > > consequence of sending extra packets.
> >
> >          I couldnt get the point of, sending extra packets during congestion.....
> > Say if a Router decides that there will be congestion ( its not in congestion now...)
> > then it decides to set the ECN bit in the IP header of certain packets in the Queue...
> > These packets should reach the receiver endpoint.... then the receiver will
> > send a ACK to the sender to indicate there is a congestion..... Doesnt this too
> > long a process.... ie., anyway a SACK is to be sent from the Receiver to the
> > sender....  In that case why not an ICMP packet from the router to the sender???
> >
> > And also, since the extra packets are going to flow in the opposite direction
> > of the actual congestion, I couldnt see any harm .... (I know this might be a
> > wrong statemnent.... but I dont have the right justification...)
> >
> > Atleast in Source Quenching method, the flooding of Data by the Sender can be
> > controlled fast than the current ECN method...right!!
> >
> > >
> > > Notice the ECN does not go towards the sender,
> > > but in the other direction. I think the intention
> > > is to have the upper layer (transport) in the
> > > receiver withhold acknowledgements to cause the
> > > sender to back off. I'm making this up, so please
> >
> >         Why at all a receiver bother about the Congetion....  I see no reasons for the
> > receiver to know about the congestion in the path.... because he cant do anything
> > about that...
> > And also by doing this by Source Quenching method.... its not mandatory that
> > both the source and destination Transport layers support the ECN capability..
> > If only sender is supporting, its more than enough... and there will be no necessity
> > to negotiate the support of this feature, end to end....
> > Router  also doesnt bother  whether an Endpoint is supporting  ECN or not...
> > he can just send an ICMP message to the sender, whether he supports ECN or not...
> > (slightly costly operation but thought its ok...)
> >
> > Thanks
> > -arun
> >
> > >
> > > correct me if I'm wrong.
> > >
> > > PMTU discovery does not face this problem. It's ok to send
> > > an extra packet back to the sender in this case.
> > >
> > > Arun Prasad wrote:
> > > >
> > > > Hi all,
> > > >     Some doubts on the procedure followed for PMTU (Path mtu) discovery and
> > > > ECN method.....
> > > >     The doubt is not related to the procedure as such, but the relation between the
> > > > two procedures stated above......
> > > > Tcp and Sctp uses ICMP message to handle the PMTU discovery procedure....
> > > > ie., the intermediate router will send an ICMP messge of "Message too Long"
> > > > to the end point sending the data packet of size greater than the MTU size of some
> > > > intermediate path to the destination.....
> > > >
> > > > For ECN, the intermediate router should set the appropriate bits in the IP header
> > > > (if it supports ECN ....) whenever there is congestion and risk of packet drops
> > > > is high....... The endpoints then correct their Congestion window accordingly and
> > > > help in reducing the Congestion in the intermediate router.....
> > > >
> > > > Why, we adopt two different methods to carry out similar (in the sense, both are
> > > > related to the Intermediate routers....)?????
> > > >     Why cant we generalise this stuff.... ie.,  the need here is some way to
> > > > communicate between the intermediate router and the endpoint.... Not sure
> > > > which of the two ways is advantageous, but cant we use the same method for
> > > > both..... ie.,
> > > > As for the PMTU discovery, the intermediate router can send an ICMP message
> > > > to the endpoint carrying a message "CONGESTION", by receiving this the
> > > > endpoint will do all appropriate actions as it does when it receives the packet set
> > > > with ECN-ECHO flag (as in present ECN procedure...)
> > > >
> > > > or vice versa ( ie., adopt the method followed by ECN  for PMTU discovery aslo,
> > > > that might be slighly tough, considering the Backward compatibility....)
> > > >
> > > >     What we can achieve by this is the simplicity  in the router implementation,
> > > > which doesnt need to do different procedure for different features of the Transport
> > > > layer..... and by maintaining this uniformity, the future extentions which demands a
> > > > similar requirement can use the same way......
> > > >
> > > > I could have missed some points....
> > > >
> > > > Any comments on this????
> > > >
> > > > Thanks
> > > > -arun
> > > > --
> > > > ****************************************************************
> > > > V.Arun Prasad
> > > > HCL Technologies Ltd.
> > > > 51, Jawaharlal Nehru Road,
> > > > Ekkattuthangal,
> > > > Guindy Industrial Estate,
> > > > Chennai - 600097.
> > > >
> > > > Contact # : 9144 - 2334174
> > > >             9144 - 2334181
> > > >             extn : 233
> > > > ****************************************************************
> > > >
> > > > _______________________________________________
> > > > tsvwg mailing list
> > > > tsvwg@ietf.org
> > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > >
> > > --
> > > // Jacob Heitz. Tel:510-747-2917
> >
> > --
> > ****************************************************************
> > V.Arun Prasad
> > HCL Technologies Ltd.
> > 51, Jawaharlal Nehru Road,
> > Ekkattuthangal,
> > Guindy Industrial Estate,
> > Chennai - 600097.
> >
> > Contact # : 9144 - 2334174
> >             9144 - 2334181
> >             extn : 233
> > ****************************************************************
>
> --
> // Jacob Heitz. Tel:510-747-2917

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr  9 16:57:04 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA22162
	for <tsvwg-archive@odin.ietf.org>; Tue, 9 Apr 2002 16:57:04 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA19798
	for tsvwg-archive@odin.ietf.org; Tue, 9 Apr 2002 16:57:05 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA17590;
	Tue, 9 Apr 2002 16:10:02 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA17562
	for <tsvwg@optimus.ietf.org>; Tue, 9 Apr 2002 16:09:56 -0400 (EDT)
Received: from auemail2.firewall.lucent.com (auemail2.lucent.com [192.11.223.163])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA20280
	for <tsvwg@ietf.org>; Tue, 9 Apr 2002 16:09:52 -0400 (EDT)
Received: from ihrh1.emsr.lucent.com (h135-1-218-53.lucent.com [135.1.218.53])
	by auemail2.firewall.lucent.com (Switch-2.1.3/Switch-2.1.0) with ESMTP id g39K8m828788;
	Tue, 9 Apr 2002 16:08:48 -0400 (EDT)
Received: from new-wopr.eng.ascend.com by ihrh1.emsr.lucent.com (8.8.8+Sun/EMS-1.5 Solaris/emsr)
	id PAA03809 for ; Tue, 9 Apr 2002 15:08:46 -0500 (CDT)
Received: from baxter.eng.ascend.com (baxter.eng.ascend.com [135.140.53.35])
	by new-wopr.eng.ascend.com (8.10.2+Sun/8.10.2) with ESMTP id g39K8kP11567;
	Tue, 9 Apr 2002 13:08:46 -0700 (PDT)
Received: from lucent.com by baxter.eng.ascend.com (8.9.3+Sun/SMI-SVR4)
	id NAA27628; Tue, 9 Apr 2002 13:08:45 -0700 (PDT)
Message-ID: <3CB3585A.7E7C670C@lucent.com>
Date: Tue, 09 Apr 2002 17:08:42 -0400
From: Srivathsa <sriva@lucent.com>
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.2.16-22 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Arun Prasad <arun@netlab.hcltech.com>
CC: tsvwg <tsvwg@ietf.org>, tcp-impl@grc.nasa.gov,
        sctp <sctpimp@netlab.hcltech.com>
References: <3CADF78E.D9E24D26@cisco.com> <3CB20DBE.3B687309@cisco.com> <3CB27092.7680BC16@netlab.hcltech.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] Re: ECN & PMTU
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Arun Prasad wrote:
> 
> Hi all,
>     Some doubts on the procedure followed for PMTU (Path mtu) discovery and
> ECN method.....
>     The doubt is not related to the procedure as such, but the relation between the
> two procedures stated above......
> Tcp and Sctp uses ICMP message to handle the PMTU discovery procedure....
> ie., the intermediate router will send an ICMP messge of "Message too Long"
> to the end point sending the data packet of size greater than the MTU size of some
> intermediate path to the destination.....
> 
> For ECN, the intermediate router should set the appropriate bits in the IP header
> (if it supports ECN ....) whenever there is congestion and risk of packet drops
> is high....... The endpoints then correct their Congestion window accordingly and
> help in reducing the Congestion in the intermediate router.....
> 
> Why, we adopt two different methods to carry out similar (in the sense, both are
> related to the Intermediate routers....)?????

Arun,
   I dont really see how the 2 situations are similar.
In the PMTU case,  the network is in normal state.

In the ECN case, there is a congestion at the router. The router cant
afford to pump more packets into the network.  In the earlier internet,
the source quench mechanism was used to combat congestion  (i.e router
sends an ICMP to the source).  In a quick evolution, routers started
using a packet-drop mechanism to combat congestion (It didnt make sense
to pump more packets into a congested network).  Now, ECN is being
recommended to combat congestion VERY EARLY. 
  i.e PMTU-like-mechanism cant be used for ECN.


>     Why cant we generalise this stuff.... ie.,  the need here is some way to
> communicate between the intermediate router and the endpoint.... Not sure
> which of the two ways is advantageous, but cant we use the same method for
> both..... ie.,
> As for the PMTU discovery, the intermediate router can send an ICMP message
> to the endpoint carrying a message "CONGESTION", by receiving this the
> endpoint will do all appropriate actions as it does when it receives the packet set
> with ECN-ECHO flag (as in present ECN procedure...)
> 
> or vice versa ( ie., adopt the method followed by ECN  for PMTU discovery aslo,
> that might be slighly tough, considering the Backward compatibility....)
> 


You cant use a ECN-like-mechanism for PMTU discovery.   As soon as an
intermediate router discovers that the packet (with DontFragment bit set
) is too large for one of its outbound interfaces,   it will have to
discard the packet.  I.e it cant relay the packet to the
destination.      In ECN scenario,  packet can still be relayed further.

-Srivathsa

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr  9 19:18:07 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA25488
	for <tsvwg-archive@odin.ietf.org>; Tue, 9 Apr 2002 19:18:07 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id TAA27817
	for tsvwg-archive@odin.ietf.org; Tue, 9 Apr 2002 19:18:07 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA27392;
	Tue, 9 Apr 2002 19:07:01 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA05170
	for <tsvwg@optimus.ietf.org>; Tue, 9 Apr 2002 13:08:55 -0400 (EDT)
Received: from goose.ehsco.com (goose.ehsco.com [207.65.203.98])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA12324
	for <tsvwg@ietf.org>; Tue, 9 Apr 2002 13:08:53 -0400 (EDT)
Received: from [68.52.3.151] (account ehall HELO ehsco.com)
  by goose.ehsco.com (CommuniGate Pro SMTP 3.5.7)
  with ESMTP-TLS id 40735; Tue, 09 Apr 2002 11:08:41 -0600
Message-ID: <3CB32023.2739D549@ehsco.com>
Date: Tue, 09 Apr 2002 12:08:52 -0500
From: "Eric A. Hall" <ehall@ehsco.com>
X-Mailer: Mozilla 4.78 [en] (Windows NT 5.0; U)
X-Accept-Language: en-US,en
MIME-Version: 1.0
To: Jacob Heitz <jheitz@lucent.com>
CC: Arun Prasad <arun@netlab.hcltech.com>, tsvwg <tsvwg@ietf.org>,
        tcp-impl@grc.nasa.gov, sctp <sctpimp@netlab.hcltech.com>
Subject: Re: [Tsvwg] ECN & PMTU
References: <3CADF78E.D9E24D26@cisco.com> <3CB20DBE.3B687309@cisco.com> <3CB27092.7680BC16@netlab.hcltech.com> <3CB285B4.9E8FBBA4@lucent.com> <3CB2BB02.21AA779C@netlab.hcltech.com> <3CB2C285.FA4806AB@lucent.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit


[I promise not to get sucked into this again, but a few clarifications are
in order.]

Jacob Heitz wrote:

> Source quenching was always controversial.
> It is considered higher impact, because it adds
> an extra packet to the network. Setting a CE
> bit adds nothing that was not there before anyway,
> so is considered to be safer, or lower impact.

...unless you also consider the TCP retransmissions to be expensive.

> Yes, it's slower than having the congested router
> send an ICMP.

> The end receiver sends an ACK anyway. Let it modify
> or withhold that ACK. This is what the receiver can
> do. It would send the ACK anyway, thus nothing
> extra is added to the network.
> 
> But consider these points:
> 
> A network is made of different makes/models of routers.
> A router may consider itself congested in all
> directions at once when it runs low on buffer memory.
> Then it may send ICMPs every which way.

Is it going to start dropping packets anyway?

> If a path is congested at many routers, each may send
> an ICMP in the reverse direction, resulting in a
> multitude of ICMPs reaching the sender.

Good point.

> When congestion collapse looms, safety first is the
> golden rule: add nothing more that was not already
> there anyway.

On half-duplex links, that is a good idea.

-- 
Eric A. Hall                                        http://www.ehsco.com/
Internet Core Protocols          http://www.oreilly.com/catalog/coreprot/


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr 10 00:44:32 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA02649
	for <tsvwg-archive@odin.ietf.org>; Wed, 10 Apr 2002 00:44:31 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id AAA14354
	for tsvwg-archive@odin.ietf.org; Wed, 10 Apr 2002 00:44:34 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id AAA13120;
	Wed, 10 Apr 2002 00:23:41 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id AAA13093
	for <tsvwg@optimus.ietf.org>; Wed, 10 Apr 2002 00:23:38 -0400 (EDT)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA01924
	for <tsvwg@ietf.org>; Wed, 10 Apr 2002 00:23:34 -0400 (EDT)
Received: from netlab.hcltech.com (arun.hclt-guindy.co.in [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id JAA00595;
	Wed, 10 Apr 2002 09:52:04 +0530
Message-ID: <3CB3BE57.2655E29C@netlab.hcltech.com>
Date: Wed, 10 Apr 2002 09:53:51 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Srivathsa <sriva@lucent.com>
CC: tsvwg <tsvwg@ietf.org>, tcp-impl@grc.nasa.gov,
        e2e <end2end-interest@postel.org>, sctp <sctpimp@netlab.hcltech.com>
References: <3CADF78E.D9E24D26@cisco.com> <3CB20DBE.3B687309@cisco.com> <3CB27092.7680BC16@netlab.hcltech.com> <3CB3585A.7E7C670C@lucent.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] Re: ECN & PMTU
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit



Srivathsa wrote:

> >     Why cant we generalise this stuff.... ie.,  the need here is some way to
> > communicate between the intermediate router and the endpoint.... Not sure
> > which of the two ways is advantageous, but cant we use the same method for
> > both..... ie.,
> > As for the PMTU discovery, the intermediate router can send an ICMP message
> > to the endpoint carrying a message "CONGESTION", by receiving this the
> > endpoint will do all appropriate actions as it does when it receives the packet set
> > with ECN-ECHO flag (as in present ECN procedure...)
> >
> > or vice versa ( ie., adopt the method followed by ECN  for PMTU discovery aslo,
> > that might be slighly tough, considering the Backward compatibility....)
> >
>
> You cant use a ECN-like-mechanism for PMTU discovery.   As soon as an
> intermediate router discovers that the packet (with DontFragment bit set
> ) is too large for one of its outbound interfaces,   it will have to
> discard the packet.  I.e it cant relay the packet to the
> destination.      In ECN scenario,  packet can still be relayed further.
>

            Ya, I agree that......  But the  method by which PMTU discovery is done
can be changed to avoid packet drops (if dropping a packet is costly... I think
yes, because it triggers Retransmissions....). We might rethink about implementing PMTU
discovery like the ECN like mechanism, with this new proposal..... The idea is

there should be  someother bit like the DF bit, say IF bit (Inform on Fragmentation) in the
IP header..... This IF bit will be set by the Transport layers if it is planning to do the PMTU
discovery.....  The router upon receiving a packet of size greater than the MTU size of the
Egress interface and the IF bit is set in the IP header... then IP layer there will go ahead
 fragmenting the packet and sending it to the destination..., but also send an ICMP
"Message too Long" message to the Sender.....  The router also sets the DF bit, and
unsets the IF bit in the IP header of the fragment packets (which ensure no more
fragmentation is done on the packet....)  ..................The sender upon receiving the ICMP
packet will just update his PMTU... Thus there is no Packet drops... so the procedure of
PMTU discovery will not  trigger any Retransmissions.... (which is expensive, because
the transport layer shrinks the Congestion window if it senses that a packe is dropped...)


Not sure whether this is already discussed or proposed etc...


Thanks
-arun

>
> -Srivathsa

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr 10 02:10:15 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA12152
	for <tsvwg-archive@odin.ietf.org>; Wed, 10 Apr 2002 02:10:15 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id CAA29409
	for tsvwg-archive@odin.ietf.org; Wed, 10 Apr 2002 02:10:18 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA28290;
	Wed, 10 Apr 2002 01:52:32 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA28259
	for <tsvwg@optimus.ietf.org>; Wed, 10 Apr 2002 01:52:29 -0400 (EDT)
Received: from sj-msg-core-1.cisco.com (sj-msg-core-1.cisco.com [171.71.163.11])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA03700
	for <tsvwg@ietf.org>; Wed, 10 Apr 2002 01:52:25 -0400 (EDT)
Received: from mira-sjcm-3.cisco.com (IDENT:mirapoint@mira-sjcm-3.cisco.com [171.69.24.15])
	by sj-msg-core-1.cisco.com (8.12.2/8.12.2) with ESMTP id g3A5pvq6000303;
	Tue, 9 Apr 2002 22:51:57 -0700 (PDT)
Received: from cisco.com (sjc-vpn1-854.cisco.com [10.21.99.86])
	by mira-sjcm-3.cisco.com (Mirapoint)
	with ESMTP id AEL26265;
	Tue, 9 Apr 2002 22:51:52 -0700 (PDT)
Message-ID: <3CB3D33B.D10D0E82@cisco.com>
Date: Tue, 09 Apr 2002 22:53:00 -0700
From: mbashyam <mbashyam@cisco.com>
X-Mailer: Mozilla 4.76 [en]C-CCK-MCD   (Windows NT 5.0; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Arun Prasad <arun@netlab.hcltech.com>
CC: Srivathsa <sriva@lucent.com>, tsvwg <tsvwg@ietf.org>,
        tcp-impl@grc.nasa.gov, e2e <end2end-interest@postel.org>,
        sctp <sctpimp@netlab.hcltech.com>
Subject: Re: [Tsvwg] Re: ECN & PMTU
References: <3CADF78E.D9E24D26@cisco.com> <3CB20DBE.3B687309@cisco.com> <3CB27092.7680BC16@netlab.hcltech.com> <3CB3585A.7E7C670C@lucent.com> <3CB3BE57.2655E29C@netlab.hcltech.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit



Arun Prasad wrote:

> Srivathsa wrote:
>
> > >     Why cant we generalise this stuff.... ie.,  the need here is some way to
> > > communicate between the intermediate router and the endpoint.... Not sure
> > > which of the two ways is advantageous, but cant we use the same method for
> > > both..... ie.,
> > > As for the PMTU discovery, the intermediate router can send an ICMP message
> > > to the endpoint carrying a message "CONGESTION", by receiving this the
> > > endpoint will do all appropriate actions as it does when it receives the packet set
> > > with ECN-ECHO flag (as in present ECN procedure...)
> > >
> > > or vice versa ( ie., adopt the method followed by ECN  for PMTU discovery aslo,
> > > that might be slighly tough, considering the Backward compatibility....)
> > >
> >
> > You cant use a ECN-like-mechanism for PMTU discovery.   As soon as an
> > intermediate router discovers that the packet (with DontFragment bit set
> > ) is too large for one of its outbound interfaces,   it will have to
> > discard the packet.  I.e it cant relay the packet to the
> > destination.      In ECN scenario,  packet can still be relayed further.
> >
>
>             Ya, I agree that......  But the  method by which PMTU discovery is done
> can be changed to avoid packet drops (if dropping a packet is costly... I think
> yes, because it triggers Retransmissions....). We might rethink about implementing PMTU
> discovery like the ECN like mechanism, with this new proposal..... The idea is
>
> there should be  someother bit like the DF bit, say IF bit (Inform on Fragmentation) in the
> IP header..... This IF bit will be set by the Transport layers if it is planning to do the PMTU
> discovery.....  The router upon receiving a packet of size greater than the MTU size of the
> Egress interface and the IF bit is set in the IP header... then IP layer there will go ahead
>  fragmenting the packet and sending it to the destination..., but also send an ICMP
> "Message too Long" message to the Sender.....  The router also sets the DF bit, and
> unsets the IF bit in the IP header of the fragment packets (which ensure no more
> fragmentation is done on the packet....)  ..................The sender upon receiving the ICMP
> packet will just update his PMTU... Thus there is no Packet drops... so the procedure of
> PMTU discovery will not  trigger any Retransmissions.... (which is expensive, because
> the transport layer shrinks the Congestion window if it senses that a packe is dropped...)

I dont understand, PMTU discovery requires that TCP go into a retransmission of all its
unacknowledged packets if there is a change in MSS.

Also in ur scheme by forwarding the fragments with DF bit on won't u  be triggering
drops in a downstream router with a smaller next hop MTU than the just announced
PMTU ?

Murali

>
>
> Not sure whether this is already discussed or proposed etc...
>
> Thanks
> -arun
>
> >
> > -Srivathsa
>
> --
> ****************************************************************
> V.Arun Prasad
> HCL Technologies Ltd.
> 51, Jawaharlal Nehru Road,
> Ekkattuthangal,
> Guindy Industrial Estate,
> Chennai - 600097.
>
> Contact # : 9144 - 2334174
>             9144 - 2334181
>             extn : 233
> ****************************************************************
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr 10 13:02:42 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA25269
	for <tsvwg-archive@odin.ietf.org>; Wed, 10 Apr 2002 13:02:37 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA05758
	for tsvwg-archive@odin.ietf.org; Wed, 10 Apr 2002 13:02:39 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA04614;
	Wed, 10 Apr 2002 12:44:33 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA05972
	for <tsvwg@optimus.ietf.org>; Wed, 10 Apr 2002 04:26:51 -0400 (EDT)
Received: from d12lmsgate.de.ibm.com (d12lmsgate.de.ibm.com [195.212.91.199])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA13715
	for <tsvwg@ietf.org>; Wed, 10 Apr 2002 04:26:46 -0400 (EDT)
Received: from d12relay02.de.ibm.com (d12relay02.de.ibm.com [9.165.215.23])
	by d12lmsgate.de.ibm.com (1.0.0) with ESMTP id KAA122808;
	Wed, 10 Apr 2002 10:25:33 +0200
Received: from etzel.zurich.ibm.com (etzel.zurich.ibm.com [9.4.64.140])
	by d12relay02.de.ibm.com (8.11.1m3/NCO/VER6.00) with SMTP id g3A8PQj157276;
	Wed, 10 Apr 2002 10:25:26 +0200
Received: from dhcp70-23.zurich.ibm.com by etzel.zurich.ibm.com (AIX 4.3/UCB 5.64/4.03)
          id AA28408 from <mwl@zurich.ibm.com>; Wed, 10 Apr 2002 10:25:21 +0200
Message-Id: <3CB3F21A.2060006@news.m.wanda.ch>
Date: Wed, 10 Apr 2002 10:04:42 +0200
From: Marcel Waldvogel <marcel@news.m.wanda.ch>
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.8) Gecko/20020204
X-Accept-Language: de, en
Mime-Version: 1.0
To: Arun Prasad <arun@netlab.hcltech.com>
Cc: Srivathsa <sriva@lucent.com>, tsvwg <tsvwg@ietf.org>,
        tcp-impl@grc.nasa.gov, e2e <end2end-interest@postel.org>,
        sctp <sctpimp@netlab.hcltech.com>
References: <3CADF78E.D9E24D26@cisco.com> <3CB20DBE.3B687309@cisco.com> <3CB27092.7680BC16@netlab.hcltech.com> <3CB3585A.7E7C670C@lucent.com> <3CB3BE57.2655E29C@netlab.hcltech.com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] Re: [e2e] ECN & PMTU
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Arun Prasad wrote:
 > Ya, I agree that...... But the method by which PMTU discovery is done
 > can be changed to avoid packet drops (if dropping a packet is
 > costly... I think yes, because it triggers Retransmissions....). We
 > might rethink about implementing PMTU discovery like the ECN like
 > mechanism, with this new proposal..... The idea is there should be
 > someother bit like the DF bit, say IF bit (Inform on Fragmentation) in
 > the IP header..... This IF bit will be set by the Transport layers if
 > it is planning to do the PMTU discovery..... The router upon receiving
 > a packet of size greater than the MTU size of the Egress interface and
 > the IF bit is set in the IP header... then IP layer there will go
 > ahead fragmenting the packet and sending it to the destination..., but
 > also send an ICMP "Message too Long" message to the Sender..... The
 > router also sets the DF bit, and unsets the IF bit in the IP header of
 > the fragment packets (which ensure no more fragmentation is done on
 > the packet....) ..................The sender upon receiving the ICMP
 > packet will just update his PMTU... Thus there is no Packet drops...
 > so the procedure of PMTU discovery will not trigger any
 > Retransmissions.... (which is expensive, because the transport layer
 > shrinks the Congestion window if it senses that a packe is dropped...)

I like the idea of an "Inform-on-Fragmentation" bit, as this would get
the PMTU right within one RTT *and* not require any retransmits. I do
not agree with your DF/IF interaction, instead I propose my own rules
(downward compatible, single RTT, no retransmit (exception below)):

* IF overrides DF.
  Any IF-capable router will ignore DF on packets that have IF set. IF
  is more powerful than DF, but DF will be needed during a transition
  period.
* Routers do not modify the IF/DF bits.
  When the sender sets IF and DF (probably the common case), there is
  no need for that.
* Routers send back an "ICMP Fragmentation needed," indicating it was
  due to IF.
  This is preferably done only on the first fragment (i.e., offset=0).
* IF-capable routers that don't want to/can't fully perform
  fragmentation (e.g., IPv6) shorten the packet into the first
  fragment.
  The "Fragmentation needed" message then contains the information
  that the packet needs to be resent.

-Marcel




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Apr 12 09:51:46 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA24813
	for <tsvwg-archive@odin.ietf.org>; Fri, 12 Apr 2002 09:51:46 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA16732
	for tsvwg-archive@odin.ietf.org; Fri, 12 Apr 2002 09:51:49 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA14537;
	Fri, 12 Apr 2002 09:06:33 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA14493
	for <tsvwg@optimus.ietf.org>; Fri, 12 Apr 2002 09:06:27 -0400 (EDT)
Received: from irmgard.exp-math.uni-essen.de (ns.exp-math.uni-essen.de [132.252.150.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA18933;
	Fri, 12 Apr 2002 09:06:23 -0400 (EDT)
Received: from biene.exp-math.uni-essen.de (biene.exp-math.uni-essen.de [132.252.150.215])
	by irmgard.exp-math.uni-essen.de (8.12.2/8.12.2) with ESMTP id g3CD6Osj027040;
	Fri, 12 Apr 2002 15:06:24 +0200
Content-Type: text/plain;
  charset="us-ascii"
From: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
Organization: University Essen
To: "'sigtran@ietf.org'" <sigtran@ietf.org>, tsvwg <tsvwg@ietf.org>
Date: Fri, 12 Apr 2002 15:06:20 +0200
X-Mailer: KMail [version 1.4]
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Message-Id: <200204121506.22219.ajung@exp-math.uni-essen.de>
Content-Transfer-Encoding: 8bit
Subject: [Tsvwg] SCTP Interop Event (AKA: B*keoff)
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Dear all,

this is just a reminder for all those that may have missed my
earlier post (and thanks to all those who have already contacted
me -- so far 8 parties with 11 implementations) :

An SCTP  interoperability event will be organized by our Group at 
the German Unversity of Essen from September 9 to September 13, 2002.
In order to arrange as early as possible for the hotel accomodations (i.e. 
reserve a block of rooms) etc., we would like to get a rough estimate of who 
(i.e. affiliation, number of persons) would be interested in participating.

So, if you consider participating in this event, please contact me
by next wednesday at the latest. Registration will start as of 
April, 29th.

For more details, please see our web page at
http://tdrwww.exp-math.uni-essen.de/inhalt/aktuell/estart.php3

Regards,

Andreas Jungmaier

- -- 
- ------------------------------------------------------------------------
Dipl.-Ing. Andreas Jungmaier              |                         _
Computer Networking Technology Group      | ASCII ribbon campaign  ( )
University of Essen                       |  - against HTML email   X
http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8ttvMXAsLBNOCSsARAl9iAKCi5fA77hjtwwMXIW9XKQAEgRcYOwCgvPkR
tq3iwyF+cqESClXyqmnr5ws=
=ml3f
-----END PGP SIGNATURE-----


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Apr 12 16:35:15 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA03376
	for <tsvwg-archive@odin.ietf.org>; Fri, 12 Apr 2002 16:35:15 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA11889
	for tsvwg-archive@odin.ietf.org; Fri, 12 Apr 2002 16:35:18 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA09017;
	Fri, 12 Apr 2002 15:55:43 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA08989
	for <tsvwg@ns.ietf.org>; Fri, 12 Apr 2002 15:55:40 -0400 (EDT)
Received: from e31.co.us.ibm.com (e31.co.us.ibm.com [32.97.110.129])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA02232
	for <tsvwg@ietf.org>; Fri, 12 Apr 2002 15:55:37 -0400 (EDT)
Received: from westrelay03.boulder.ibm.com (westrelay03.boulder.ibm.com [9.17.194.24])
	by e31.co.us.ibm.com (8.12.2/8.12.2) with ESMTP id g3CJtcsp041598
	for <tsvwg@ietf.org>; Fri, 12 Apr 2002 15:55:39 -0400
Received: from schemer.endicott.ibm.com (schemer.endicott.ibm.com [9.60.11.237])
	by westrelay03.boulder.ibm.com (8.11.1m3/NCO/VER6.1) with ESMTP id g3CJtU899098;
	Fri, 12 Apr 2002 13:55:30 -0600
Received: by schemer.endicott.ibm.com (Postfix, from userid 1000)
	id E61911A53E; Fri, 12 Apr 2002 15:55:28 -0400 (EDT)
From: Will Stockdell <willstoc@schemer.endicott.ibm.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <15543.15280.770436.872789@schemer.endicott.ibm.com>
Date: Fri, 12 Apr 2002 15:55:28 -0400
To: tsvwg@ietf.org
X-Mailer: VM 7.00 under 21.4 (patch 6) "Common Lisp" XEmacs Lucid
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] ECN and congestion avoidance
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

I am preparing to implement ECN and in reading the RFC (3168) I have
come across something puzzling.  In section 6.1.2 it says:

"The indication of congestion should be treated just as a congestion
loss in  non-ECN-Capable TCP. That is, the TCP source halves the
congestion window "cwnd" and reduces the slow start threshold
"ssthresh"."

According to RFC 2581:
"Furthermore, upon a [retransmit] timeout cwnd MUST be set to no more
than the loss window, LW, which equals 1 full-sized segment
(regardless of the value of IW)."

So what am I missing?  Should the congestion window be halved or set
to one full-size segment when congestion is experienced?  Am I
misreading something?

Thanks,

Will




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Apr 12 22:13:32 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA09275
	for <tsvwg-archive@odin.ietf.org>; Fri, 12 Apr 2002 22:13:32 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id WAA26800
	for tsvwg-archive@odin.ietf.org; Fri, 12 Apr 2002 22:13:34 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id VAA25921;
	Fri, 12 Apr 2002 21:44:36 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id VAA25891
	for <tsvwg@ns.ietf.org>; Fri, 12 Apr 2002 21:44:34 -0400 (EDT)
Received: from poplar.cs.washington.edu (mail@poplar.cs.washington.edu [128.95.2.24])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA09104
	for <tsvwg@ietf.org>; Fri, 12 Apr 2002 21:44:29 -0400 (EDT)
Received: from nspring by poplar.cs.washington.edu with local (Exim 3.33 #1 (Debian))
	id 16wCaa-0005K1-00; Fri, 12 Apr 2002 18:44:28 -0700
Date: Fri, 12 Apr 2002 18:44:28 -0700
From: Neil Spring <nspring@cs.washington.edu>
To: Will Stockdell <willstoc@schemer.endicott.ibm.com>
Cc: tsvwg@ietf.org
Subject: Re: [Tsvwg] ECN and congestion avoidance
Message-ID: <20020413014419.GA20196@cs.washington.edu>
References: <15543.15280.770436.872789@schemer.endicott.ibm.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <15543.15280.770436.872789@schemer.endicott.ibm.com>
User-Agent: Mutt/1.3.28i
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

On Fri, Apr 12, 2002 at 03:55:28PM -0400, Will Stockdell wrote:
> I am preparing to implement ECN and in reading the RFC (3168) I have
> come across something puzzling.  In section 6.1.2 it says:
> 
> "The indication of congestion should be treated just as a congestion
> loss in  non-ECN-Capable TCP. That is, the TCP source halves the
> congestion window "cwnd" and reduces the slow start threshold
> "ssthresh"."
> 
> According to RFC 2581:
> "Furthermore, upon a [retransmit] timeout cwnd MUST be set to no more
> than the loss window, LW, which equals 1 full-sized segment
> (regardless of the value of IW)."
> 
> So what am I missing?  Should the congestion window be halved or set
> to one full-size segment when congestion is experienced?  Am I
> misreading something?

Will,

Without looking up the context, I read those passages as:

1) ECN signals are equivalent to packet losses for
congestion: the window is halved in fast retransmit /
fast recovery because data is still flowing.

2) Timeout is handled normally, where the window is reduced
to 1 to reestablish ack pacing.

I think you're assuming 'timeout <==> congestion' when
'timeout -> congestion' (3-dup ack -> congestion too).
Or did I misread your confusion?  I assume the right way
for a Tahoe-TCP (which lacks fast retranmission) to respond
to ECE would be cwnd-halving.

hope it helps,
-neil

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Apr 16 04:10:10 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA26165
	for <tsvwg-archive@odin.ietf.org>; Tue, 16 Apr 2002 04:10:09 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id EAA26894
	for tsvwg-archive@odin.ietf.org; Tue, 16 Apr 2002 04:10:12 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA25325;
	Tue, 16 Apr 2002 03:33:58 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA25286
	for <tsvwg@ns.ietf.org>; Tue, 16 Apr 2002 03:33:55 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-207.dallas.net [209.44.42.207])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA25743
	for <tsvwg@ietf.org>; Tue, 16 Apr 2002 03:33:49 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id CAA05978
	for tsvwg@ietf.org; Tue, 16 Apr 2002 02:33:51 -0500
Date: Tue, 16 Apr 2002 02:33:51 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: tsvwg@ietf.org
Message-ID: <20020416023351.A5729@openss7.org>
Reply-To: bidulock@openss7.org
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2i
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Subject: [Tsvwg] GPL'd STREAMS SCTP Implemntation for Linux 2.4
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

fyi


I have just released a Linux 2.4.x SMP kernel version of the RFC
2960 SCTP implementation based on LiS 2.13.17 (Linux STREAMS).
Whereas the release of last Nov. only ran on Linux 2.2.x, this
release has been verified to run on Linux 2.4.18 SMP.  The
release provides an OpenGroup TPI interface as well as an NPI
interface.

The implementation can be downloaded from:

	http://www.oepnss7.org/new/download.html

Included is the ability to set SCTP protocol parameter values way
outside the recommended values (for example, an RTO-Min of zero
:) and has an adaptive multi-homed routing mechanisms which has
been demonstrated to exhibit 5 ms fail-over times between
destinations.

Also included is a modified Nagle algorithm for SCTP which
provides adaptive bundling of data chunks to greatly reduce the
number of small packets and increase network efficiency for
signalling applications that send many small messages (such as
SIGTRAN UAs).

Included are test programs which demonstrate fail-over and
throughput with 2 out of 3 destination failures under 40% noise
and which also conformance test the interface.

Also available under GPL is an implementation of SIGTRAN
draft-ietf-sigtran-m2pa-04 which conforms to the 76 conformance
test cases documented in draft-bidulock-sigtran-m2pa-test-01.

A Solaris 8 port should be available soon.

--brian

-- 
Brian F. G. Bidulock
bidulock@openss7.org
http://www.openss7.org/

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 16 07:55:03 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA29928
	for <tsvwg-archive@odin.ietf.org>; Tue, 16 Apr 2002 07:55:03 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA07787
	for tsvwg-archive@odin.ietf.org; Tue, 16 Apr 2002 07:55:06 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA06303;
	Tue, 16 Apr 2002 07:30:50 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA11351
	for <tsvwg@optimus.ietf.org>; Fri, 12 Apr 2002 08:03:34 -0400 (EDT)
Received: from CNRI.Reston.VA.US (localhost [127.0.0.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA12795;
	Fri, 12 Apr 2002 08:03:30 -0400 (EDT)
Message-Id: <200204121203.IAA12795@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: IETF-Announce: ;
Cc: tsvwg@ietf.org
From: Internet-Drafts@ietf.org
Reply-to: Internet-Drafts@ietf.org
Date: Fri, 12 Apr 2002 08:03:29 -0400
Subject: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-initwin-03.txt
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the Transport Area Working Group Working Group of the IETF.

	Title		: Increasing TCP's Initial Window
	Author(s)	: M. Allman, S. Floyd, C. Partridge
	Filename	: draft-ietf-tsvwg-initwin-03.txt
	Pages		: 13
	Date		: 11-Apr-02
	
This document specifies an optional standard for TCP to increase the
permitted initial window from one segment to roughly 4K bytes,
replacing RFC 2414.  This document discusses the advantages and
disadvantages of the higher initial window.  The document includes
discussion of experiments and simulations showing that the higher
initial window does not lead to congestion collapse. Finally, the
document provides guidance on implementation issues.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-initwin-03.txt

To remove yourself from the IETF Announcement list, send a message to 
ietf-announce-request with the word unsubscribe in the body of the message.

Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-tsvwg-initwin-03.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-tsvwg-initwin-03.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<20020411133027.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-tsvwg-initwin-03.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-tsvwg-initwin-03.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<20020411133027.I-D@ietf.org>

--OtherAccess--

--NextPart--




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 16 07:56:22 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA29960
	for <tsvwg-archive@odin.ietf.org>; Tue, 16 Apr 2002 07:56:22 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA07814
	for tsvwg-archive@odin.ietf.org; Tue, 16 Apr 2002 07:56:25 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA06781;
	Tue, 16 Apr 2002 07:35:47 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA06752
	for <tsvwg@optimus.ietf.org>; Tue, 16 Apr 2002 07:35:44 -0400 (EDT)
Received: from gorilla.mchh.siemens.de (gorilla.mchh.siemens.de [194.138.158.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA29468
	for <tsvwg@ietf.org>; Tue, 16 Apr 2002 07:35:39 -0400 (EDT)
Received: from blues.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.227])
	by gorilla.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id NAA20224;
	Tue, 16 Apr 2002 13:35:34 +0200 (MET DST)
Received: from mchh273e.demchh201e.icn.siemens.de ([139.21.200.83])
	by blues.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id NAA28167;
	Tue, 16 Apr 2002 13:35:05 +0200 (MET DST)
Received: by MCHH273E with Internet Mail Service (5.5.2653.19)
	id <268PB7NQ>; Tue, 16 Apr 2002 13:35:34 +0200
Message-ID: <F92978223B01D311ACFF0008C71EE19D015DF7AF@MCHH225E>
From: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
To: "'Poetzl, Joachim'" <Joachim.Poetzl@telekom.de>, tsvwg@ietf.org
Subject: RE: [Tsvwg] SCTP and NAT
Date: Tue, 16 Apr 2002 13:35:33 +0200
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id HAA06753
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Hello,

please see my comments in-line.

Best regards
Michael

Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	Poetzl, Joachim [SMTP:Joachim.Poetzl@telekom.de]
> Sent:	Thursday, April 11, 2002 4:44 PM
> To:	tsvwg@ietf.org
> Subject:	[Tsvwg] SCTP and NAT
> 
> Hi SCTP experts,
> 
> who can help?
> 
> Does SCTP work with NAT? 
> What are the related drafts and RFCs?
> Any experience on that topic?
	[Tuexen Michael]  You may want to look into the very new RFC
	http://www.ietf.org/rfc/rfc3257.txt

> What are advantages / disadvantages using SCTP in conjunction with TLS or IPSec?
> 
> What about the discussion on the TLS over SCTP Draft (draft-ietf-tsvwg-tls-over-sctp-00.txt)? The draft is still in version 0.
	[Tuexen Michael]  It is still in version 00, but has completed the WG last call and the IESG last call just started today.

> Thanks for your help in advance.
> 
> Regards
> 
> > Joachim P> ötzl
> > 
> ........................................................................................
> Deutsche Telekom AG
> Zentralbereich Netzinfrastruktur (ZB NI)
> NI196a13; Signalling, Gateways and Switching Systems
> 64307 DARMSTADT, GERMANY 
> Phone:  +49 6151 83-2134
> Fax:      +49 6151 83-4577
> email:	joachim.poetzl@telekom.de
> ........................................................................................
> 
> 
> 
> 
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 16 07:57:08 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA29995
	for <tsvwg-archive@odin.ietf.org>; Tue, 16 Apr 2002 07:57:08 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA07839
	for tsvwg-archive@odin.ietf.org; Tue, 16 Apr 2002 07:57:11 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA06071;
	Tue, 16 Apr 2002 07:30:04 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA28452
	for <tsvwg@optimus.ietf.org>; Thu, 11 Apr 2002 20:37:43 -0400 (EDT)
Received: from CNRI.Reston.VA.US (localhost [127.0.0.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA21740;
	Thu, 11 Apr 2002 20:37:39 -0400 (EDT)
Message-Id: <200204120037.UAA21740@ietf.org>
To: IETF-Announce: ;
Cc: tsvwg@ietf.org, ietf-tls@lists.certicom.com
From: The IESG <iesg-secretary@ietf.org>
Reply-to: iesg@ietf.org
Date: Thu, 11 Apr 2002 20:37:38 -0400
Subject: [Tsvwg] Last Call: TLS over SCTP to Proposed Standard
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


The IESG has received a request from the Transport Area Working Group
Working Group to consider TLS over SCTP
<draft-ietf-tsvwg-tls-over-sctp-00.txt> as a Proposed Standard.

The IESG plans to make a decision in the next few weeks, and solicits
final comments on this action.  Please send any comments to the
iesg@ietf.org or ietf@ietf.org mailing lists by April 25, 2002.

Files can be obtained via
http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-tls-over-sctp-00.txt




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 16 07:58:03 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA00025
	for <tsvwg-archive@odin.ietf.org>; Tue, 16 Apr 2002 07:58:03 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA07878
	for tsvwg-archive@odin.ietf.org; Tue, 16 Apr 2002 07:58:05 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA05988;
	Tue, 16 Apr 2002 07:29:50 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA20229
	for <tsvwg@optimus.ietf.org>; Thu, 11 Apr 2002 10:44:36 -0400 (EDT)
Received: from mail1.telekom.de (mail1.telekom.de [62.225.183.235])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA28535
	for <tsvwg@ietf.org>; Thu, 11 Apr 2002 10:44:31 -0400 (EDT)
Received: from g8pbr.blf01.telekom.de by G8SBV.dmz.telekom.de with ESMTP for tsvwg@ietf.org; Thu, 11 Apr 2002 16:42:42 +0200
Received: by G8PBR.blf01.telekom.de with Internet Mail Service (5.5.2653.19)
	id <2W5603F0>; Thu, 11 Apr 2002 16:44:00 +0200
Message-Id: <A456C746A5E2D2118FC70090270F673B02FC1615@U8P19.blf01.telekom.de>
From: "Poetzl, Joachim" <Joachim.Poetzl@telekom.de>
To: tsvwg@ietf.org
Date: Thu, 11 Apr 2002 16:43:56 +0200
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id KAA20230
Subject: [Tsvwg] SCTP and NAT
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Hi SCTP experts,

who can help?

Does SCTP work with NAT? 
What are the related drafts and RFCs?
Any experience on that topic?

What are advantages / disadvantages using SCTP in conjunction with TLS or IPSec?

What about the discussion on the TLS over SCTP Draft (draft-ietf-tsvwg-tls-over-sctp-00.txt)? The draft is still in version 0.

Thanks for your help in advance.

Regards

> Joachim Pötzl
> 
........................................................................................
Deutsche Telekom AG
Zentralbereich Netzinfrastruktur (ZB NI)
NI196a13; Signalling, Gateways and Switching Systems
64307 DARMSTADT, GERMANY 
Phone:  +49 6151 83-2134
Fax:      +49 6151 83-4577
email:	joachim.poetzl@telekom.de
........................................................................................





_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Apr 16 09:25:38 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA02484
	for <tsvwg-archive@odin.ietf.org>; Tue, 16 Apr 2002 09:25:38 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA13942
	for tsvwg-archive@odin.ietf.org; Tue, 16 Apr 2002 09:25:41 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA11508;
	Tue, 16 Apr 2002 08:58:21 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAB11475
	for <tsvwg@ns.ietf.org>; Tue, 16 Apr 2002 08:58:17 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA01241
	for <tsvwg@ietf.org>; Tue, 16 Apr 2002 08:58:13 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3GCvuH26345;
	Tue, 16 Apr 2002 07:57:56 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CBC1FD3.DF6222B3@stewart.chicago.il.us>
Date: Tue, 16 Apr 2002 07:57:56 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: "Poetzl, Joachim" <Joachim.Poetzl@telekom.de>
CC: tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP and NAT
References: <A456C746A5E2D2118FC70090270F673B02FC1615@U8P19.blf01.telekom.de>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

A small comment or two..

"Poetzl, Joachim" wrote:

> Does SCTP work with NAT?

I think Michael as pointed out a "new" RFC that discusses this in
some detail. But I will add a few rambling thoughts on the matter....

NAT's WILL work with SCTP if you are talking singly homed and the
SCTP implementation as enough address selection sense to NOT include
any addresses in the address list. I believe our implementation on
BSD will do this (available at http://www.sctp.org and soon it
will be released as part of kame). The basic idea is that the 
SCTP implementation behind the NAT must realize it has a private
address and then only send either the INIT or INIT-ACK back to
the "global" address without listing any of its addresses.

If the SCTP implementation is multi-homed and would like to 
be able to take advantage of that ... say something like:

           (Private Address)-----NAT-1-----IP-X(Global address)
    IP-P1-----+
SCTP
    IP-P2-----+
           (Private Address)-----NAT-2-----IP-Y(Global address)

The SCTP implementation would require assistance from the two
NAT's in order to make this work..

You would need the NAT's to:

1) Do stateful inspection of the address lists and substitute
    IP-P1/IP-P2 for IP-X and IP-X (in the INIT/INIT-ACK messages). 
2) Communicate with the peer NAT through some mechanism to
   allocate a COMMON port number and setup the mapping.
3) Coordinate the tear down of this state when the association
   ends.. note that EITHER side could get the tear down indication.

I don't believe there is ANY way this can be done today without
further work on NAT's which I DO NOT advocate. A far better way
to get multi-homing is to move to IPv6 and then both sides can
have nice global addresses :>

R

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Apr 16 09:39:37 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA03545
	for <tsvwg-archive@odin.ietf.org>; Tue, 16 Apr 2002 09:39:36 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA15773
	for tsvwg-archive@odin.ietf.org; Tue, 16 Apr 2002 09:39:40 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA12623;
	Tue, 16 Apr 2002 09:10:51 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA12596
	for <tsvwg@ns.ietf.org>; Tue, 16 Apr 2002 09:10:48 -0400 (EDT)
Received: from mgw-x1.nokia.com (mgw-x1.nokia.com [131.228.20.21])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA01516
	for <tsvwg@ietf.org>; Tue, 16 Apr 2002 09:10:39 -0400 (EDT)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir05nok.ntc.nokia.com (esvir05nokt.ntc.nokia.com [172.21.143.37])
	by mgw-x1.nokia.com (Switch-2.2.1/Switch-2.2.0) with ESMTP id g3GD9mu10214
	for <tsvwg@ietf.org>; Tue, 16 Apr 2002 16:09:48 +0300 (EET DST)
Received: from esebh001.NOE.Nokia.com (unverified) by esvir05nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5a4bf58e7dac158f25076@esvir05nok.ntc.nokia.com>;
 Tue, 16 Apr 2002 16:10:40 +0300
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh001.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Tue, 16 Apr 2002 16:10:40 +0300
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] SCTP and NAT
Date: Tue, 16 Apr 2002 16:10:40 +0300
Message-ID: <58AEDE27D48F884285007ABF827FF26308EF87@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] SCTP and NAT
Thread-Index: AcHlRmzaMDOWQbhmTBSeVfCLWSAHYgAAEMWQ
To: <randall@stewart.chicago.il.us>, <Joachim.Poetzl@telekom.de>
Cc: <tsvwg@ietf.org>
X-OriginalArrivalTime: 16 Apr 2002 13:10:40.0554 (UTC) FILETIME=[1B8A50A0:01C1E548]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id JAA12597
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi all!

	I'm not a real expert about NATs and I would like people to comment about this mail, but... what about the Hostname Address Parameter?

	It is true that in this case you need a DNS aware of the existence of the NAT (I think they are called Application Level Gateway in RFC 2663, I haven't read that new RFC Michael pointed out), but they do exist (otherwise, how would you set up a server inside a stub domain?)... The receiver of such parameter would do the DNS query and should get the globally unique addresses of the peer... The NATs in its normal behavior will do the rest...

	This was the reason behind the use of the Hostname Address Parameter, right? Or am I missing something?

	Thanks! Any comment is very welcome!

	BR Iván Arias Rodríguez

> -----Original Message-----
> From: ext Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: 16. April 2002 15:58
> To: Poetzl, Joachim
> Cc: tsvwg@ietf.org
> Subject: Re: [Tsvwg] SCTP and NAT
> 
> 
> A small comment or two..
> 
> "Poetzl, Joachim" wrote:
> 
> > Does SCTP work with NAT?
> 
> I think Michael as pointed out a "new" RFC that discusses this in
> some detail. But I will add a few rambling thoughts on the matter....
> 
> NAT's WILL work with SCTP if you are talking singly homed and the
> SCTP implementation as enough address selection sense to NOT include
> any addresses in the address list. I believe our implementation on
> BSD will do this (available at http://www.sctp.org and soon it
> will be released as part of kame). The basic idea is that the 
> SCTP implementation behind the NAT must realize it has a private
> address and then only send either the INIT or INIT-ACK back to
> the "global" address without listing any of its addresses.
> 
> If the SCTP implementation is multi-homed and would like to 
> be able to take advantage of that ... say something like:
> 
>            (Private Address)-----NAT-1-----IP-X(Global address)
>     IP-P1-----+
> SCTP
>     IP-P2-----+
>            (Private Address)-----NAT-2-----IP-Y(Global address)
> 
> The SCTP implementation would require assistance from the two
> NAT's in order to make this work..
> 
> You would need the NAT's to:
> 
> 1) Do stateful inspection of the address lists and substitute
>     IP-P1/IP-P2 for IP-X and IP-X (in the INIT/INIT-ACK messages). 
> 2) Communicate with the peer NAT through some mechanism to
>    allocate a COMMON port number and setup the mapping.
> 3) Coordinate the tear down of this state when the association
>    ends.. note that EITHER side could get the tear down indication.
> 
> I don't believe there is ANY way this can be done today without
> further work on NAT's which I DO NOT advocate. A far better way
> to get multi-homing is to move to IPv6 and then both sides can
> have nice global addresses :>
> 
> R
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Apr 16 09:48:20 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA04032
	for <tsvwg-archive@odin.ietf.org>; Tue, 16 Apr 2002 09:48:20 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA16417
	for tsvwg-archive@odin.ietf.org; Tue, 16 Apr 2002 09:48:24 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA14094;
	Tue, 16 Apr 2002 09:27:04 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA14065
	for <tsvwg@ns.ietf.org>; Tue, 16 Apr 2002 09:27:00 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA02577
	for <tsvwg@ietf.org>; Tue, 16 Apr 2002 09:26:55 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3GDQaH26398;
	Tue, 16 Apr 2002 08:26:36 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CBC268B.10C75535@stewart.chicago.il.us>
Date: Tue, 16 Apr 2002 08:26:35 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: Joachim.Poetzl@telekom.de, tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP and NAT
References: <58AEDE27D48F884285007ABF827FF26308EF87@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Ivan:

Yes, I do agree. The host-name-address parameter was designed for JUST
such a thing. You do need DNS aware NAT's. But more importantly you
need an SCTP implementation that understands and sends this
parameter. For a USER land implementation this is not to hard of
issue ... since you are in a library someplace, in user space, and
making a hostname lookup call is NOT an issue.

But when you think in terms of a KERNEL implementation the issue
becomes MUCH MUCH harder. You must from within the kernel make
an UP call to a user application (aka DNS). Then suspend the process
doing the INIT (if we are in a blocking mode) and wait for the response
from user land to the kernel. This gets really ugly really fast.. it
can be done.. don't get me wrong... but it is a difficult thing to do.

One other little item ... if you really want to have multi-homing you
still need some NAT-NAT coordination protocol to allocate a shared
PORT amongst the two NAT's that you are multi-homed with. Without this
feature in a NAT I just don't see the value in writting all this kernel
code.. since if you just send a INIT/INIT-ACK off without listing
addresses you can setup the association singly-homed... 

Thus the only thing that I can think the DNS work (upcall in kernel..
uggh)
will do for you is help you when you need to be multi-homed with a NAT
and the NAT-NAT allocation issue is still left open.. so I just don't
see how this feature does anything to help you...IMHO...

But of course I never was very excited about the host-name-address
parameter 
:-0

R




Ivan.Arias-Rodriguez@nokia.com wrote:
> 
>         Hi all!
> 
>         I'm not a real expert about NATs and I would like people to comment about this mail, but... what about the Hostname Address Parameter?
> 
>         It is true that in this case you need a DNS aware of the existence of the NAT (I think they are called Application Level Gateway in RFC 2663, I haven't read that new RFC Michael pointed out), but they do exist (otherwise, how would you set up a server inside a stub domain?)... The receiver of such parameter would do the DNS query and should get the globally unique addresses of the peer... The NATs in its normal behavior will do the rest...
> 
>         This was the reason behind the use of the Hostname Address Parameter, right? Or am I missing something?
> 
>         Thanks! Any comment is very welcome!
> 
>         BR Iván Arias Rodríguez
> 
> > -----Original Message-----
> > From: ext Randall Stewart [mailto:randall@stewart.chicago.il.us]
> > Sent: 16. April 2002 15:58
> > To: Poetzl, Joachim
> > Cc: tsvwg@ietf.org
> > Subject: Re: [Tsvwg] SCTP and NAT
> >
> >
> > A small comment or two..
> >
> > "Poetzl, Joachim" wrote:
> >
> > > Does SCTP work with NAT?
> >
> > I think Michael as pointed out a "new" RFC that discusses this in
> > some detail. But I will add a few rambling thoughts on the matter....
> >
> > NAT's WILL work with SCTP if you are talking singly homed and the
> > SCTP implementation as enough address selection sense to NOT include
> > any addresses in the address list. I believe our implementation on
> > BSD will do this (available at http://www.sctp.org and soon it
> > will be released as part of kame). The basic idea is that the
> > SCTP implementation behind the NAT must realize it has a private
> > address and then only send either the INIT or INIT-ACK back to
> > the "global" address without listing any of its addresses.
> >
> > If the SCTP implementation is multi-homed and would like to
> > be able to take advantage of that ... say something like:
> >
> >            (Private Address)-----NAT-1-----IP-X(Global address)
> >     IP-P1-----+
> > SCTP
> >     IP-P2-----+
> >            (Private Address)-----NAT-2-----IP-Y(Global address)
> >
> > The SCTP implementation would require assistance from the two
> > NAT's in order to make this work..
> >
> > You would need the NAT's to:
> >
> > 1) Do stateful inspection of the address lists and substitute
> >     IP-P1/IP-P2 for IP-X and IP-X (in the INIT/INIT-ACK messages).
> > 2) Communicate with the peer NAT through some mechanism to
> >    allocate a COMMON port number and setup the mapping.
> > 3) Coordinate the tear down of this state when the association
> >    ends.. note that EITHER side could get the tear down indication.
> >
> > I don't believe there is ANY way this can be done today without
> > further work on NAT's which I DO NOT advocate. A far better way
> > to get multi-homing is to move to IPv6 and then both sides can
> > have nice global addresses :>
> >
> > R
> >
> > --
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> >

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Apr 16 10:17:20 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA05731
	for <tsvwg-archive@odin.ietf.org>; Tue, 16 Apr 2002 10:17:20 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA18920
	for tsvwg-archive@odin.ietf.org; Tue, 16 Apr 2002 10:17:24 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA16732;
	Tue, 16 Apr 2002 09:51:00 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA16702
	for <tsvwg@ns.ietf.org>; Tue, 16 Apr 2002 09:50:57 -0400 (EDT)
Received: from hrtades9.atea.be (viap111.atea.be [194.78.143.111])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA04182
	for <tsvwg@ietf.org>; Tue, 16 Apr 2002 09:50:50 -0400 (EDT)
Received: from hrtades10.atea.be (siemens.atea.be [139.10.143.141]) by hrtades9.atea.be with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2653.13)
	id 2ZLS7JGY; Tue, 16 Apr 2002 15:50:56 +0200
Received: by siemens.atea.be with Internet Mail Service (5.5.2653.19)
	id <GYDMSQ98>; Tue, 16 Apr 2002 15:51:01 +0200
Message-ID: <E76F715C0429D5118F2100508BB9EDEE037F5A02@hrtades7.atea.be>
From: Coene Lode <Lode.Coene@siemens.atea.be>
To: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        "Poetzl, Joachim"
	 <Joachim.Poetzl@telekom.de>
Cc: tsvwg@ietf.org
Subject: RE: [Tsvwg] SCTP and NAT
Date: Tue, 16 Apr 2002 15:46:05 +0200
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

A few comments below:

> -----Original Message-----
> From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: Tuesday, April 16, 2002 2:58 PM
> To: Poetzl, Joachim
> Cc: tsvwg@ietf.org
> Subject: Re: [Tsvwg] SCTP and NAT
> 
> 
> A small comment or two..
> 
> "Poetzl, Joachim" wrote:
> 
> > Does SCTP work with NAT?
> 
> I think Michael as pointed out a "new" RFC that discusses this in
> some detail. But I will add a few rambling thoughts on the matter....

Another draft, originally part of the "new" RFC 3257 describes this a bit in
more detail.
see http://search.ietf.org/internet-drafts/draft-coene-sctp-multihome-03.txt
It got a bit lost but I am trying to put it back on the track for
informational

> 
> NAT's WILL work with SCTP if you are talking singly homed and the
> SCTP implementation as enough address selection sense to NOT include
> any addresses in the address list. I believe our implementation on
> BSD will do this (available at http://www.sctp.org and soon it
> will be released as part of kame). The basic idea is that the 
> SCTP implementation behind the NAT must realize it has a private
> address and then only send either the INIT or INIT-ACK back to
> the "global" address without listing any of its addresses.
> 
> If the SCTP implementation is multi-homed and would like to 
> be able to take advantage of that ... say something like:
> 
>            (Private Address)-----NAT-1-----IP-X(Global address)
>     IP-P1-----+
> SCTP
>     IP-P2-----+
>            (Private Address)-----NAT-2-----IP-Y(Global address)
> 
> The SCTP implementation would require assistance from the two
> NAT's in order to make this work..
> 
> You would need the NAT's to:
> 
> 1) Do stateful inspection of the address lists and substitute
>     IP-P1/IP-P2 for IP-X and IP-X (in the INIT/INIT-ACK messages). 
> 2) Communicate with the peer NAT through some mechanism to
>    allocate a COMMON port number and setup the mapping.
> 3) Coordinate the tear down of this state when the association
>    ends.. note that EITHER side could get the tear down indication.
> 
> I don't believe there is ANY way this can be done today without
> further work on NAT's which I DO NOT advocate. A far better way
> to get multi-homing is to move to IPv6 and then both sides can
> have nice global addresses :>
> 
> R
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 

Agree with problem description. 
Solutions:
INIT, INIT-ACK and data traffic goes through a single NAT box and then you
get a single point-of-failure. Not what you are looking for...
or
you do some NAT-NAT communication coupled together with some SCTP-NAT
communication(and that should only be the simple case where only one of the
endpoint is behind a NAT, I wouldn't even dream of multiple NAT box crossing
in-between) 
Thus the solution would be invariably complex.

Therefore the simplest solution is to have NO NAT between your endpoints.
- own private IPv4 network(but doesn't solve the requirement below)
- need interconnecting with a different operator, well then IPv6 springs to
mind.

yours sincerly,
Lode 

Siemens atea
atealaan 34          2200 Herentals, Belgium
E-mail: lode.coene@siemens.atea.be
Tel: +32-14-252081
Fax: +32-14-253212


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Apr 16 17:15:24 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA20556
	for <tsvwg-archive@odin.ietf.org>; Tue, 16 Apr 2002 17:15:23 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA15147
	for tsvwg-archive@odin.ietf.org; Tue, 16 Apr 2002 17:15:27 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA12859;
	Tue, 16 Apr 2002 16:32:29 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA12831
	for <tsvwg@ns.ietf.org>; Tue, 16 Apr 2002 16:32:27 -0400 (EDT)
Received: from sbcs.cs.sunysb.edu (sbcs.cs.sunysb.edu [130.245.1.15])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA15410
	for <tsvwg@ietf.org>; Tue, 16 Apr 2002 16:32:23 -0400 (EDT)
Received: from compserv3 (compserv3 [130.245.1.116])
	by sbcs.cs.sunysb.edu (8.9.3/8.9.3) with ESMTP id QAA07533;
	Tue, 16 Apr 2002 16:28:22 -0400 (EDT)
Date: Tue, 16 Apr 2002 16:28:24 -0400 (EDT)
From: Kostas Pentikousis <kostas@cs.sunysb.edu>
X-X-Sender:  <kostas@compserv3>
To: Jacob Heitz <jheitz@lucent.com>
cc: Arun Prasad <arun@netlab.hcltech.com>, tsvwg <tsvwg@ietf.org>,
        <tcp-impl@grc.nasa.gov>
Subject: Re: [Tsvwg] ECN & PMTU
In-Reply-To: <3CB2C285.FA4806AB@lucent.com>
Message-ID: <Pine.GSO.4.33.0204161612310.17719-100000@compserv3>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

On Tue, 9 Apr 2002, Jacob Heitz wrote:

[...]
  | Source quenching was always controversial.
  | It is considered higher impact, because it adds
  | an extra packet to the network. Setting a CE
  | bit adds nothing that was not there before anyway,
  | so is considered to be safer, or lower impact.
  |
  | Yes, it's slower than having the congested router
  | send an ICMP.

I'm not sure that's really true. In fact, I've heard several times that Source
Quench (SQ) via ICMP is more expensive for current generation routers than
setting a bit. I remember reading that any packets needing "special" treatment
are placed in a separate (lower priority) queue and are dealt with when the
router load drops (a bit). Setting a bit is not considered "special" treatment
but creating a packet (setting source and destination addresses, etc.) is. If
this is true, and ECN was spec'ed so that the router would send an SQ ICMP
packet back to the sender instead of setting a bit (RFC 3168), then in most
cases the congestion notification would actually arrive later than an RTT.

Just my $0.02,

Kostas
______________________________________________________________________
Kostas Pentikousis @ CS Stony Brook @ http://www.cs.sunysb.edu/~kostas


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr 17 02:20:45 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA11234
	for <tsvwg-archive@odin.ietf.org>; Wed, 17 Apr 2002 02:20:45 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id CAA18258
	for tsvwg-archive@odin.ietf.org; Wed, 17 Apr 2002 02:20:47 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id CAA17349;
	Wed, 17 Apr 2002 02:00:59 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id CAA17315
	for <tsvwg@optimus.ietf.org>; Wed, 17 Apr 2002 02:00:57 -0400 (EDT)
Received: from cwcsun41.cwc.nus.edu.sg (cwcsun41.cwc.nus.edu.sg [137.132.163.102])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA03864
	for <tsvwg@ietf.org>; Wed, 17 Apr 2002 02:00:51 -0400 (EDT)
Received: from liawys.cwc.nus.edu.sg (8273000520.cwc.nus.edu.sg [172.16.1.171])
	by cwcsun41.cwc.nus.edu.sg (8.9.3/8.9.3) with ESMTP id NAA04042
	for <tsvwg@ietf.org>; Wed, 17 Apr 2002 13:59:30 +0800 (SGT)
Message-Id: <4.3.2.7.2.20020417131407.00ad87c0@postman.cwc.nus.edu.sg>
X-Sender: liawys@postman.cwc.nus.edu.sg
X-Mailer: QUALCOMM Windows Eudora Version 4.3.2
Date: Wed, 17 Apr 2002 14:08:56 +0800
To: tsvwg@ietf.org
From: Liaw Yong Shyang <liawys@cwc.nus.edu.sg>
Subject: [Tsvwg] ECN-based TCP
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi,

I have been looking at the problem of TCP over lossy wireless network for 
sometimes, and I would appreciate if you can comment on the following.  If 
you need more information, please feel free to contact me.  Thank you.

=====================

The main problem of TCP over wireless is that it cannot differentiate the 
type of losses (i.e. congestive or bad channel), and TCP indiscriminately 
invokes congestion control whenever losses occur.  Since ECN provides a 
mechanism for network routers to signal congestion to the end hosts, TCP 
can rely on ECN as congestion indication (i.e. ECN-based TCP), thereby is 
able to differentiate the type of losses.  This allows TCP to reduce its 
sending rate only when ECN indicates congestion during Fast Retransmit, and 
avoid unnecessary reduction.  I have simulation results to show that this 
greatly enhances TCP performance over wireless link under various network 
conditions.

Of course, the above assumes that all the network routers support ECN, and 
this is not realistic.  I expect that over the next 10 - 15 years, ECN will 
be widely implemented in the network router and deployed in the 
network.  However, we can expect that over this period of gradual 
deployment, there will be some old routers that are not ECN-enabled in the 
network.  Therefore, there is a need to determine whether all the routers 
along the path taken by a TCP connection is ECN-enabled.  And ECN-based TCP 
(i.e. only ECN is taken a congestion signal) is used only when the path is 
ECN-enable, otherwise losses shall be taken as sign of congestion as in the 
traditional TCP.  I would like to extend ECN to provide this capability, so 
that TCP can negotiate whether to rely on ECN only, or  both ECN and losses 
as congestion signal.  This can be done by providing a 8-bit "ECN-counter" 
field in the IP header (i.e. as IP option).  The "ECN-counter" shall be 
treated just the same as the "Time-to-live" field in the IP header -- 
decrement by one on passing each ECN-enable router along the path.  On 
reaching the destinated end-host, if "ecn-counter" equals the 
"time-to-live" field, then we can conclude the path is ECN-enabled, and the 
end-host will inform the TCP sender.


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr 17 07:22:10 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA14584
	for <tsvwg-archive@odin.ietf.org>; Wed, 17 Apr 2002 07:22:09 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA02357
	for tsvwg-archive@odin.ietf.org; Wed, 17 Apr 2002 07:22:12 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA01227;
	Wed, 17 Apr 2002 06:59:52 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA01202
	for <tsvwg@optimus.ietf.org>; Wed, 17 Apr 2002 06:59:49 -0400 (EDT)
Received: from prue.eim.surrey.ac.uk (IDENT:exim@prue.eim.surrey.ac.uk [131.227.76.5])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA14199
	for <tsvwg@ietf.org>; Wed, 17 Apr 2002 06:59:46 -0400 (EDT)
Received: from phaestos.ee.surrey.ac.uk ([131.227.88.14] ident=eep1lw)
	by prue.eim.surrey.ac.uk with esmtp (Exim 3.33 #4)
	id 16xnA1-0002hE-00; Wed, 17 Apr 2002 11:59:37 +0100
Date: Wed, 17 Apr 2002 11:59:32 +0100 (BST)
From: Lloyd Wood <l.wood@eim.surrey.ac.uk>
X-X-Sender: eep1lw@phaestos.ee.surrey.ac.uk
Reply-To: Lloyd Wood <L.Wood@eim.surrey.ac.uk>
To: Liaw Yong Shyang <liawys@cwc.nus.edu.sg>
cc: tsvwg@ietf.org
Subject: Re: [Tsvwg] ECN-based TCP
In-Reply-To: <4.3.2.7.2.20020417131407.00ad87c0@postman.cwc.nus.edu.sg>
Message-ID: <Pine.SOL.4.43.0204171149370.293-100000@phaestos.ee.surrey.ac.uk>
Organization: speaking for none
X-url: http://www.ee.surrey.ac.uk/Personal/L.Wood/
X-no-archive: yes
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
X-Scanner: exiscan *16xnA1-0002hE-00*z.07A63Xdn6* (SECM, UniS)
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

On Wed, 17 Apr 2002, Liaw Yong Shyang wrote:

> I have been looking at the problem of TCP over lossy wireless network for
> sometimes, and I would appreciate if you can comment on the following.  If
> you need more information, please feel free to contact me.  Thank you.
>
> =====================
>
> The main problem of TCP over wireless is that it cannot differentiate the
> type of losses (i.e. congestive or bad channel), and TCP indiscriminately
> invokes congestion control whenever losses occur.  Since ECN provides a
> mechanism for network routers to signal congestion to the end hosts, TCP
> can rely on ECN as congestion indication (i.e. ECN-based TCP), thereby is
> able to differentiate the type of losses.  This allows TCP to reduce its
> sending rate only when ECN indicates congestion during Fast Retransmit, and
> avoid unnecessary reduction.

ECN is an imperfect measure of congestion; packets marked with an ECN
bit are themselves subject to congestion. If you receive a packet
marked with an ECN bit (and it really is an ECN bit rather than some
old-style QoS marking...) you know there's congestion in your path,
but if you don't you can't be sure (because the packet might be lost).
So, TCP can never expect to rely only on ECN. (In IPv6 there's no
header checksum -- so can you really trust the ECN bit you see?)

The ECN counter as an IP header option would require *large* amounts
of router processing, as packets with options are not handled on a
fast path or in hardware. And if path rerouting occurs during a
TCP session or you have split paths...

TCP negotiation's in RFC3168 section 6; splits paths are further down.

cheers,

L.

>  I have simulation results to show that this
> greatly enhances TCP performance over wireless link under various network
> conditions.
>
> Of course, the above assumes that all the network routers support ECN, and
> this is not realistic.  I expect that over the next 10 - 15 years, ECN will
> be widely implemented in the network router and deployed in the
> network.  However, we can expect that over this period of gradual
> deployment, there will be some old routers that are not ECN-enabled in the
> network.  Therefore, there is a need to determine whether all the routers
> along the path taken by a TCP connection is ECN-enabled.  And ECN-based TCP
> (i.e. only ECN is taken a congestion signal) is used only when the path is
> ECN-enable, otherwise losses shall be taken as sign of congestion as in the
> traditional TCP.  I would like to extend ECN to provide this capability, so
> that TCP can negotiate whether to rely on ECN only, or  both ECN and losses
> as congestion signal.  This can be done by providing a 8-bit "ECN-counter"
> field in the IP header (i.e. as IP option).  The "ECN-counter" shall be
> treated just the same as the "Time-to-live" field in the IP header --
> decrement by one on passing each ECN-enable router along the path.  On
> reaching the destinated end-host, if "ecn-counter" equals the
> "time-to-live" field, then we can conclude the path is ECN-enabled, and the
> end-host will inform the TCP sender.
>
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
>

<L.Wood@surrey.ac.uk>PGP<http://www.ee.surrey.ac.uk/Personal/L.Wood/>


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr 17 08:05:16 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA15197
	for <tsvwg-archive@odin.ietf.org>; Wed, 17 Apr 2002 08:05:16 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA04534
	for tsvwg-archive@odin.ietf.org; Wed, 17 Apr 2002 08:05:18 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA03394;
	Wed, 17 Apr 2002 07:42:15 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA03367
	for <tsvwg@optimus.ietf.org>; Wed, 17 Apr 2002 07:42:11 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA14805
	for <tsvwg@ietf.org>; Wed, 17 Apr 2002 07:42:07 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3HBfWH28778;
	Wed, 17 Apr 2002 06:41:32 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CBD5F6B.9525B45F@stewart.chicago.il.us>
Date: Wed, 17 Apr 2002 06:41:32 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Liaw Yong Shyang <liawys@cwc.nus.edu.sg>, Sally Floyd <floyd@aciri.org>
CC: tsvwg@ietf.org
Subject: Re: [Tsvwg] ECN-based TCP
References: <4.3.2.7.2.20020417131407.00ad87c0@postman.cwc.nus.edu.sg>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Liaw:

A thought for you to consider... just because a router
is capable of ECN does NOT mean it will always ONLY use
ECN to indicate congestion. A router will keep different
thresholds... the first one will be where it starts marking
packets... but if congestion continues it will start dropping
packets..

Basically you can NOT rely on ECN by itself so that you can
tell the difference between a drop due to error and a drop
due to congestion... IMO... Sally what do you think?

R

Liaw Yong Shyang wrote:
> 
> Hi,
> 
> I have been looking at the problem of TCP over lossy wireless network for
> sometimes, and I would appreciate if you can comment on the following.  If
> you need more information, please feel free to contact me.  Thank you.
> 
> =====================
> 
> The main problem of TCP over wireless is that it cannot differentiate the
> type of losses (i.e. congestive or bad channel), and TCP indiscriminately
> invokes congestion control whenever losses occur.  Since ECN provides a
> mechanism for network routers to signal congestion to the end hosts, TCP
> can rely on ECN as congestion indication (i.e. ECN-based TCP), thereby is
> able to differentiate the type of losses.  This allows TCP to reduce its
> sending rate only when ECN indicates congestion during Fast Retransmit, and
> avoid unnecessary reduction.  I have simulation results to show that this
> greatly enhances TCP performance over wireless link under various network
> conditions.
> 
> Of course, the above assumes that all the network routers support ECN, and
> this is not realistic.  I expect that over the next 10 - 15 years, ECN will
> be widely implemented in the network router and deployed in the
> network.  However, we can expect that over this period of gradual
> deployment, there will be some old routers that are not ECN-enabled in the
> network.  Therefore, there is a need to determine whether all the routers
> along the path taken by a TCP connection is ECN-enabled.  And ECN-based TCP
> (i.e. only ECN is taken a congestion signal) is used only when the path is
> ECN-enable, otherwise losses shall be taken as sign of congestion as in the
> traditional TCP.  I would like to extend ECN to provide this capability, so
> that TCP can negotiate whether to rely on ECN only, or  both ECN and losses
> as congestion signal.  This can be done by providing a 8-bit "ECN-counter"
> field in the IP header (i.e. as IP option).  The "ECN-counter" shall be
> treated just the same as the "Time-to-live" field in the IP header --
> decrement by one on passing each ECN-enable router along the path.  On
> reaching the destinated end-host, if "ecn-counter" equals the
> "time-to-live" field, then we can conclude the path is ECN-enabled, and the
> end-host will inform the TCP sender.
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr 17 12:53:22 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA26745
	for <tsvwg-archive@odin.ietf.org>; Wed, 17 Apr 2002 12:53:22 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id MAA29416
	for tsvwg-archive@odin.ietf.org; Wed, 17 Apr 2002 12:53:24 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA27105;
	Wed, 17 Apr 2002 12:28:39 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA11291
	for <tsvwg@optimus.ietf.org>; Wed, 17 Apr 2002 11:58:29 -0400 (EDT)
Received: from elk.icir.org (elk.icir.org [192.150.187.21])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA24282
	for <tsvwg@ietf.org>; Wed, 17 Apr 2002 11:58:24 -0400 (EDT)
Received: from elk.icir.org (localhost [127.0.0.1])
	by elk.icir.org (8.11.6/8.11.3) with ESMTP id g3HFwO197247;
	Wed, 17 Apr 2002 08:58:24 -0700 (PDT)
	(envelope-from floyd@elk.icir.org)
Message-Id: <200204171558.g3HFwO197247@elk.icir.org>
To: Randall Stewart <randall@stewart.chicago.il.us>
cc: tsvwg@ietf.org, Liaw Yong Shyang <liawys@cwc.nus.edu.sg>
From: Sally Floyd <floyd@icir.org>
Subject: Re: [Tsvwg] ECN-based TCP 
Date: Wed, 17 Apr 2002 08:58:24 -0700
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

>Basically you can NOT rely on ECN by itself so that you can
>tell the difference between a drop due to error and a drop
>due to congestion... IMO... Sally what do you think?

Yep. 

This is stated in the ECN document, RFC3168, in Section 19.1:

"The CE codepoint of the ECN field is only used by routers as an
indication of congestion during periods of *moderate* congestion.
ECN-capable routers should drop rather than mark packets during heavy
congestion even if the router's queue is not yet full."

and is discussed in my short paper on "A Report on Some Recent
Developments in TCP Congestion Control", available from
"http://www.icir.org/floyd/papers.html".

- Sally


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Wed Apr 17 23:21:40 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA16290
	for <tsvwg-archive@odin.ietf.org>; Wed, 17 Apr 2002 23:21:39 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id XAA08027
	for tsvwg-archive@odin.ietf.org; Wed, 17 Apr 2002 23:21:43 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id WAA06660;
	Wed, 17 Apr 2002 22:55:52 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id WAA06629
	for <tsvwg@ns.ietf.org>; Wed, 17 Apr 2002 22:55:49 -0400 (EDT)
Received: from cwcsun41.cwc.nus.edu.sg (cwcsun41.cwc.nus.edu.sg [137.132.163.102])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA15956
	for <tsvwg@ietf.org>; Wed, 17 Apr 2002 22:55:43 -0400 (EDT)
Received: from liawys.cwc.nus.edu.sg (8273000520.cwc.nus.edu.sg [172.16.1.171])
	by cwcsun41.cwc.nus.edu.sg (8.9.3/8.9.3) with ESMTP id KAA09929;
	Thu, 18 Apr 2002 10:54:25 +0800 (SGT)
Message-Id: <4.3.2.7.2.20020418101113.00b21c90@postman.cwc.nus.edu.sg>
X-Sender: liawys@postman.cwc.nus.edu.sg
X-Mailer: QUALCOMM Windows Eudora Version 4.3.2
Date: Thu, 18 Apr 2002 11:03:25 +0800
To: Lloyd Wood <L.Wood@eim.surrey.ac.uk>
From: Liaw Yong Shyang <liawys@cwc.nus.edu.sg>
Subject: Re: [Tsvwg] ECN-based TCP
Cc: tsvwg@ietf.org
In-Reply-To: <Pine.SOL.4.43.0204171149370.293-100000@phaestos.ee.surrey.
 ac.uk>
References: <4.3.2.7.2.20020417131407.00ad87c0@postman.cwc.nus.edu.sg>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

At 11:59 AM 4/17/02 +0100, Lloyd Wood wrote:

>ECN is an imperfect measure of congestion; packets marked with an ECN
>bit are themselves subject to congestion. If you receive a packet
>marked with an ECN bit (and it really is an ECN bit rather than some
>old-style QoS marking...) you know there's congestion in your path,
>but if you don't you can't be sure (because the packet might be lost).

Yes, packets can be lost due to buffer over-flow, or along the path to 
destination.
To make ECN more robust to losses, we can do the followings:
a. mark more packets (i.e. higher marking probability near the maxthresh) 
-- this shall aid against marked packet loss along the path.
b. use timeout as congestion indication (as a backup), and reset cwnd to 
1  (as in traditional TCP). -- this shall help to capture persistent 
congestion in the router.

With occasional losses, I think it is not a problem.

>So, TCP can never expect to rely only on ECN. (In IPv6 there's no
>header checksum -- so can you really trust the ECN bit you see?)

Yes this is a problem, and I haven't really think much about the security 
aspect.  What about the "ECN-nonce" proposed in the Internet draft "Robust 
ECN Signaling with Nonces" at 
http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-tcp-nonce-02.txt?  Can 
that help?


>The ECN counter as an IP header option would require *large* amounts
>of router processing, as packets with options are not handled on a
>fast path or in hardware. And if path rerouting occurs during a
>TCP session or you have split paths...

Yes, IP option is not a nice thing, but there is no more space in the IP 
header.  For IPv6, it can be an IPv6 header extension.  Actually, it is not 
necessary to send "ecn-counter" option in every packets all the time.  We 
may send the option on packets for the 1st few windows to determine that 
the paths taken (including multi-paths) are ECN-capable, and then every few 
other windows regularly -- assuming that paths will not changes too frequently.


>TCP negotiation's in RFC3168 section 6; splits paths are further down.

OK, I will take a look at it.  Thank.


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Thu Apr 18 02:06:12 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA26491
	for <tsvwg-archive@odin.ietf.org>; Thu, 18 Apr 2002 02:06:12 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id CAA26819
	for tsvwg-archive@odin.ietf.org; Thu, 18 Apr 2002 02:06:14 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA25563;
	Thu, 18 Apr 2002 01:45:23 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA25534
	for <tsvwg@ns.ietf.org>; Thu, 18 Apr 2002 01:45:20 -0400 (EDT)
Received: from cwcsun41.cwc.nus.edu.sg (cwcsun41.cwc.nus.edu.sg [137.132.163.102])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA18191
	for <tsvwg@ietf.org>; Thu, 18 Apr 2002 01:45:11 -0400 (EDT)
Received: from liawys.cwc.nus.edu.sg (8273000520.cwc.nus.edu.sg [172.16.1.171])
	by cwcsun41.cwc.nus.edu.sg (8.9.3/8.9.3) with ESMTP id NAA26365;
	Thu, 18 Apr 2002 13:42:50 +0800 (SGT)
Message-Id: <4.3.2.7.2.20020418110452.00ae0eb0@postman.cwc.nus.edu.sg>
X-Sender: liawys@postman.cwc.nus.edu.sg
X-Mailer: QUALCOMM Windows Eudora Version 4.3.2
Date: Thu, 18 Apr 2002 13:51:45 +0800
To: Randall Stewart <randall@stewart.chicago.il.us>,
        Sally Floyd <floyd@aciri.org>
From: Liaw Yong Shyang <liawys@cwc.nus.edu.sg>
Subject: Re: [Tsvwg] ECN-based TCP
Cc: tsvwg@ietf.org
In-Reply-To: <3CBD5F6B.9525B45F@stewart.chicago.il.us>
References: <4.3.2.7.2.20020417131407.00ad87c0@postman.cwc.nus.edu.sg>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi Randall & Sally,

Thank for your reply.

In the current ECN (RFC 3168), RED AQM will drop packet once it exceeds 
maxthresh -- this is designed to use loss as an indication of 
congestion.  In the case where ECN is to be used as the only congestion 
indication, the packets shall be marked instead of being dropped.  Of 
course when the buffer is full, packets will still be dropped.

I think, to take ECN  as the only congestion indication, there must have 
some supports from the router':
- mark-tail, instead of drop tail (for ECT packets)
- provide "sufficient" buffer so that TCP packets are marked once beyond 
maxthresh
- perhaps, mark more packets (higher marking probability) especially when 
it is near maxthresh, may be the Adaptive RED -- help to be robust against 
loss along the path.

Additionally, TCP must also have timeout mechanism as a backup, in case 
ECN-marked packet are lost along the path, and buffer-overflow in router.

I have done some work on this, and have written up a paper (accepted for 
WOC2002) -- I am on the process of setting up my web-page, and will inform 
you once it is up (soon).  In this, I showed that it is possible to keep 
ECN-based TCP not more "aggressive" than traditional TCP/SACK during 
congestion.

Further improvements may be done.  For example, fine-tuning the level of 
congestion according to the amount of marked packets received.  Say if 50% 
of the packets received are marked, then TCP sender shall consider more 
drastic action then just halve the cwnd (i.e. reset cwnd to 1).

Thanks.


At 06:41 AM 4/17/02 -0500, Randall Stewart wrote:
>Liaw:
>
>A thought for you to consider... just because a router
>is capable of ECN does NOT mean it will always ONLY use
>ECN to indicate congestion. A router will keep different
>thresholds... the first one will be where it starts marking
>packets... but if congestion continues it will start dropping
>packets..
>
>Basically you can NOT rely on ECN by itself so that you can
>tell the difference between a drop due to error and a drop
>due to congestion... IMO... Sally what do you think?



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Apr 18 03:30:41 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA27967
	for <tsvwg-archive@odin.ietf.org>; Thu, 18 Apr 2002 03:30:41 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id DAA02676
	for tsvwg-archive@odin.ietf.org; Thu, 18 Apr 2002 03:30:44 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA00842;
	Thu, 18 Apr 2002 03:08:36 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA00772
	for <tsvwg@optimus.ietf.org>; Thu, 18 Apr 2002 03:08:30 -0400 (EDT)
Received: from cwcsun41.cwc.nus.edu.sg (cwcsun41.cwc.nus.edu.sg [137.132.163.102])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA27593
	for <tsvwg@ietf.org>; Thu, 18 Apr 2002 03:08:24 -0400 (EDT)
Received: from liawys.cwc.nus.edu.sg (8273000520.cwc.nus.edu.sg [172.16.1.171])
	by cwcsun41.cwc.nus.edu.sg (8.9.3/8.9.3) with ESMTP id PAA04460;
	Thu, 18 Apr 2002 15:07:05 +0800 (SGT)
Message-Id: <4.3.2.7.2.20020418151253.00b177d0@postman.cwc.nus.edu.sg>
X-Sender: liawys@postman.cwc.nus.edu.sg
X-Mailer: QUALCOMM Windows Eudora Version 4.3.2
Date: Thu, 18 Apr 2002 15:15:58 +0800
To: Randall Stewart <randall@stewart.chicago.il.us>,
        Sally Floyd <floyd@aciri.org>
From: Liaw Yong Shyang <liawys@cwc.nus.edu.sg>
Subject: Re: [Tsvwg] ECN-based TCP
Cc: tsvwg@ietf.org
In-Reply-To: <3CBD5F6B.9525B45F@stewart.chicago.il.us>
References: <4.3.2.7.2.20020417131407.00ad87c0@postman.cwc.nus.edu.sg>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi Randall & Sally,

My paper  "Decoupling of TCP Congestion Control and Loss Recovery for 
Wireless Network" can be found at http://www.cwc.nus.edu.sg/~liawys/.  Thanks.


At 06:41 AM 4/17/02 -0500, Randall Stewart wrote:
>Liaw:
>
>A thought for you to consider... just because a router
>is capable of ECN does NOT mean it will always ONLY use
>ECN to indicate congestion. A router will keep different
>thresholds... the first one will be where it starts marking
>packets... but if congestion continues it will start dropping
>packets..
>
>Basically you can NOT rely on ECN by itself so that you can
>tell the difference between a drop due to error and a drop
>due to congestion... IMO... Sally what do you think?


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Apr 18 06:25:13 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA00489
	for <tsvwg-archive@odin.ietf.org>; Thu, 18 Apr 2002 06:25:12 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA14246
	for tsvwg-archive@odin.ietf.org; Thu, 18 Apr 2002 06:25:14 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA13023;
	Thu, 18 Apr 2002 06:00:50 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA12991
	for <tsvwg@optimus.ietf.org>; Thu, 18 Apr 2002 06:00:46 -0400 (EDT)
Received: from relay3.ee.ncu.edu.tw (relay3.ee.ncu.edu.tw [140.115.70.23])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA00286
	for <tsvwg@ietf.org>; Thu, 18 Apr 2002 06:00:39 -0400 (EDT)
Received: from sophia (wm23.csie.ncu.edu.tw [140.115.155.123])
	(authenticated (0 bits))
	by relay3.ee.ncu.edu.tw (8.11.6/8.11.6) with ESMTP id g3I9x1h14320;
	Thu, 18 Apr 2002 17:59:01 +0800 (CST)
Message-ID: <003001c1e6bf$9acf3f60$7b9b738c@wmlab.csie.ncu.edu.tw>
From: "Elia Chen" <lamb@wmlab.csie.ncu.edu.tw>
To: "Lloyd Wood" <L.Wood@eim.surrey.ac.uk>,
        "Liaw Yong Shyang" <liawys@cwc.nus.edu.sg>
Cc: <tsvwg@ietf.org>
References: <4.3.2.7.2.20020417131407.00ad87c0@postman.cwc.nus.edu.sg> <4.3.2.7.2.20020418101113.00b21c90@postman.cwc.nus.edu.sg>
Subject: Re: [Tsvwg] ECN-based TCP
Date: Thu, 18 Apr 2002 17:58:34 +0800
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 6.00.2600.0000
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit


----- Original Message ----- 
From: "Liaw Yong Shyang" <liawys@cwc.nus.edu.sg>
To: "Lloyd Wood" <L.Wood@eim.surrey.ac.uk>
Cc: <tsvwg@ietf.org>
Sent: Thursday, April 18, 2002 11:03 AM
Subject: Re: [Tsvwg] ECN-based TCP


> At 11:59 AM 4/17/02 +0100, Lloyd Wood wrote:
> 
> >ECN is an imperfect measure of congestion; packets marked with an ECN
> >bit are themselves subject to congestion. If you receive a packet
> >marked with an ECN bit (and it really is an ECN bit rather than some
> >old-style QoS marking...) you know there's congestion in your path,
> >but if you don't you can't be sure (because the packet might be lost).
> 
> Yes, packets can be lost due to buffer over-flow, or along the path to 
> destination.
> To make ECN more robust to losses, we can do the followings:
> a. mark more packets (i.e. higher marking probability near the maxthresh) 
> -- this shall aid against marked packet loss along the path.
> b. use timeout as congestion indication (as a backup), and reset cwnd to 
> 1  (as in traditional TCP). -- this shall help to capture persistent 
> congestion in the router.
> 
> With occasional losses, I think it is not a problem.

Hello Shyang,

Your solution b is still a question for improving the perfomance issue in
the error-prone link environment. Under the situation of both frequent 
drops and occassional errors to packets, what should the sender react?

Indeed, only ECN is not perfect, as Lloyd said. The reason is on the
difficulty to distinguish path congested or not congested, in the situation 
mentioned above. While receiver got less packets, it got less ECN bits.

Randall and Sally just reminded that packet losses is still the primary 
signal for congestion. However, this conclusion is not a good news. >_<

I've ever seen a special approach to apply the ECN mark to wireless
networking. It uses the context by ECN marks and loss as the congestion
coherence. You can find it at Raj Jain's web page:

http://www.cis.ohio-state.edu/~jain/papers.html

Wish it can do some help to you, and good luck. ^_^


--
Shi-Yang Elia Chen <lamb@wmlab.csie.ncu.edu.tw>



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Apr 18 10:03:18 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA06550
	for <tsvwg-archive@odin.ietf.org>; Thu, 18 Apr 2002 10:03:18 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA28452
	for tsvwg-archive@odin.ietf.org; Thu, 18 Apr 2002 10:03:21 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA26647;
	Thu, 18 Apr 2002 09:39:04 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA26614
	for <tsvwg@optimus.ietf.org>; Thu, 18 Apr 2002 09:39:00 -0400 (EDT)
Received: from chuckie.dgms.com (angelica.ulticom.com [208.255.120.2])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA05624
	for <tsvwg@ietf.org>; Thu, 18 Apr 2002 09:38:55 -0400 (EDT)
Received: from pcmiller (localhost [127.0.0.1])
	by chuckie.dgms.com (8.9.3/8.9.3) with SMTP id JAA01429
	for <tsvwg@ietf.org>; Thu, 18 Apr 2002 09:38:27 -0400 (EDT)
From: "Stuart Miller" <stuart.miller@ulticom.com>
To: <tsvwg@ietf.org>
Date: Thu, 18 Apr 2002 09:38:34 -0400
Message-ID: <004301c1e6de$57432170$f22119ac@pcmiller>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook CWS, Build 9.0.2416 (9.0.2911.0)
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Importance: Normal
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] sctp-ref-impl 4.05 on Solaris 8
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi,

I am trying to get the sctp-ref-impl version 4.05 working on a Solaris 8
Sparc machine.
I have 2 ethernet cards, hme0 and hme1.  I have IPv4 and IPv6 addresses
configured
as follows:

    lo0: flags=1000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4> mtu 8232 index 1
            inet 127.0.0.1 netmask ff000000
    hme0: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index
2
            inet 172.25.2.56 netmask ffff0000 broadcast 172.25.255.255
    hme1: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index
3
            inet 10.2.56.1 netmask ffffff00 broadcast 10.2.56.255
    lo0: flags=2000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv6> mtu 8252 index 1
            inet6 ::1/128
    hme0: flags=2000841<UP,RUNNING,MULTICAST,IPv6> mtu 1500 index 2
            inet6 fe80::a00:20ff:feb9:e9d/10
    hme0:1: flags=2000841<UP,RUNNING,MULTICAST,IPv6> mtu 1500 index 2
            inet6 fec0::1:a00:20ff:feb9:e9d/64
    hme1: flags=2000841<UP,RUNNING,MULTICAST,IPv6> mtu 1500 index 3
            inet6 fe80::a00:20ff:feb9:e9d/10
    hme1:1: flags=2000841<UP,RUNNING,MULTICAST,IPv6> mtu 1500 index 3
            inet6 fec0::2:a00:20ff:feb9:e9d/64

When I start "sctpDeamon" only the IPv4 addresses are displayed.

I repeated this on a Linux RedHat 7.1 machine, and the IPv4 and IPv6
addresses
are displayed.

Has anyone used the reference implementation yet on Solaris with IPv6
addresses?

Thank you,
Stu Miller


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Thu Apr 18 22:53:53 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA00346
	for <tsvwg-archive@odin.ietf.org>; Thu, 18 Apr 2002 22:53:53 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id WAA22139
	for tsvwg-archive@odin.ietf.org; Thu, 18 Apr 2002 22:53:56 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id WAA21585;
	Thu, 18 Apr 2002 22:33:23 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id WAA21556
	for <tsvwg@ns.ietf.org>; Thu, 18 Apr 2002 22:33:19 -0400 (EDT)
Received: from cwcsun41.cwc.nus.edu.sg (cwcsun41.cwc.nus.edu.sg [137.132.163.102])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA29945
	for <tsvwg@ietf.org>; Thu, 18 Apr 2002 22:33:14 -0400 (EDT)
Received: from liawys.cwc.nus.edu.sg (8273000520.cwc.nus.edu.sg [172.16.1.171])
	by cwcsun41.cwc.nus.edu.sg (8.9.3/8.9.3) with ESMTP id KAA25342;
	Fri, 19 Apr 2002 10:31:45 +0800 (SGT)
Message-Id: <4.3.2.7.2.20020419091532.00adc2c0@postman.cwc.nus.edu.sg>
X-Sender: liawys@postman.cwc.nus.edu.sg
X-Mailer: QUALCOMM Windows Eudora Version 4.3.2
Date: Fri, 19 Apr 2002 10:40:05 +0800
To: "Elia Chen" <lamb@wmlab.csie.ncu.edu.tw>,
        "Lloyd Wood" <L.Wood@eim.surrey.ac.uk>
From: Liaw Yong Shyang <liawys@cwc.nus.edu.sg>
Subject: Re: [Tsvwg] ECN-based TCP
Cc: <tsvwg@ietf.org>
In-Reply-To: <003001c1e6bf$9acf3f60$7b9b738c@wmlab.csie.ncu.edu.tw>
References: <4.3.2.7.2.20020417131407.00ad87c0@postman.cwc.nus.edu.sg>
 <4.3.2.7.2.20020418101113.00b21c90@postman.cwc.nus.edu.sg>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi Elia,

> >
> > Yes, packets can be lost due to buffer over-flow, or along the path to
> > destination.
> > To make ECN more robust to losses, we can do the followings:
> > a. mark more packets (i.e. higher marking probability near the maxthresh)
> > -- this shall aid against marked packet loss along the path.
> > b. use timeout as congestion indication (as a backup), and reset cwnd to
> > 1  (as in traditional TCP). -- this shall help to capture persistent
> > congestion in the router.
> >
> > With occasional losses, I think it is not a problem.
>
>Hello Shyang,
>
>Your solution b is still a question for improving the perfomance issue in
>the error-prone link environment. Under the situation of both frequent
>drops and occassional errors to packets, what should the sender react?

First, let me thank for your comment.

Solution b (timeout) is not mean to improve the performance, but to capture 
the persistent congestion and invoke congestion control since congestion 
(buffer overflowing) at the router will drop packets.  So performance is 
improved only when there is bit-error losses, and network is not heavily 
congested.  When packets are dropped due to congestion, the sender will 
retransmit and go into fast recovery mode.  There are 3 possible cases here:
a) if the congestion persists, the retransmit pkt will be dropped, and 
eventually TCP will timeout, and reset cwnd to 1.
b) if the congestion subsides, but still above maxthreash, the retransmit 
pkt will be marked, and hence eventually sender will invoke congestion control.
c) if the congestion has subsides, and hence the packet is not marked, then 
TCP sender shall continue from where it was before the pkt was dropped.


>Indeed, only ECN is not perfect, as Lloyd said. The reason is on the
>difficulty to distinguish path congested or not congested, in the situation
>mentioned above. While receiver got less packets, it got less ECN bits.
>
>Randall and Sally just reminded that packet losses is still the primary
>signal for congestion. However, this conclusion is not a good news. >_<

I agree with you and the rest (Lloyd, Randall and Sally) that I cannot rely 
on ECN as the only congestion indication.  This is especially so when the 
network is heavily congested, and when the link BER is very "high", which 
will result in substantially increase in the packet loss rate.  Perhaps, it 
is OK to rely on ECN only when network is not heavily congested (i.e. the 
loss rate is below certain threshold), and when TCP receiver sees a 
significant increase in packet loss, it shall indicate ECE in its ACK to 
the sender.


>I've ever seen a special approach to apply the ECN mark to wireless
>networking. It uses the context by ECN marks and loss as the congestion
>coherence. You can find it at Raj Jain's web page:

Would it have the same problem, as packets may just get dropped instead, 
and marked packet may be lost along the path?




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Apr 19 02:02:40 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA06746
	for <tsvwg-archive@odin.ietf.org>; Fri, 19 Apr 2002 02:02:40 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id CAA14199
	for tsvwg-archive@odin.ietf.org; Fri, 19 Apr 2002 02:02:41 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA13177;
	Fri, 19 Apr 2002 01:47:58 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA13145
	for <tsvwg@ns.ietf.org>; Fri, 19 Apr 2002 01:47:55 -0400 (EDT)
Received: from hermes.fm.intel.com (fmr01.intel.com [192.55.52.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA02933
	for <tsvwg@ietf.org>; Fri, 19 Apr 2002 01:47:53 -0400 (EDT)
Received: from talaria.fm.intel.com (talaria.fm.intel.com [10.1.192.39])
	by hermes.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.39 2002/04/15 17:47:23 root Exp $) with ESMTP id g3J5mwL20541
	for <tsvwg@ietf.org>; Fri, 19 Apr 2002 05:48:58 GMT
Received: from fmsmsxvs040.fm.intel.com (fmsmsxv040-1.fm.intel.com [132.233.48.108])
	by talaria.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.16 2002/04/15 17:47:00 root Exp $) with SMTP id g3J5nh706511
	for <tsvwg@ietf.org>; Fri, 19 Apr 2002 05:49:44 GMT
Received: from fmsmsx28.fm.intel.com ([132.233.42.28])
 by fmsmsxvs040.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002041822471206283
 ; Thu, 18 Apr 2002 22:47:13 -0700
Received: by fmsmsx28.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <2KDWAHW8>; Thu, 18 Apr 2002 22:47:50 -0700
Message-ID: <A1171C2850D8D4119D7B0002A508C76402A42A9A@bgsmsx91.iind.intel.com>
From: "Buraga, Partha S" <p_buraga@trillium.com>
To: "'tsvwg@ietf.org'" <tsvwg@ietf.org>
Cc: "'rrs@cisco.com'" <rrs@cisco.com>,
        "'Arun Prasad'"<arun@netlab.hcltech.com>,
        "'Ivan.Arias-Rodriguez@nokia.com'"<Ivan.Arias-Rodriguez@nokia.com>,
        "'Andreas Jungmaier'" <ajung@exp-math.uni-essen.de>
Date: Thu, 18 Apr 2002 22:45:37 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Subject: [Tsvwg] HEARTBEAT failure
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi

In multihoming scenario,  a node has got three addresses x,y,z. In which,  x
-- address ( NIC ) was failed.

Application ( user of SCTP ) specifies the source address -- x,  for an
outgoing HEARTBEAT chunk. Tho address -- x was failed, kernel deliver the
packet to the destination thro interface y or z. 

But, when the HEARTBEAT - ACK responded by the peer, lands at interface --
x, it will fail. And after ' n ' number of retries, Association will fail. 

If this is the case, how can we establish association with the peer. I feel,
this problem will araise when we specify source address to the outgoing
chunk. If not, With IP_ANY, kernel can choose the address and association
could be established.

Can you please look into this and reply back.

thanks in advance
partha

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Apr 19 02:45:35 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA12009
	for <tsvwg-archive@odin.ietf.org>; Fri, 19 Apr 2002 02:45:35 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id CAA16561
	for tsvwg-archive@odin.ietf.org; Fri, 19 Apr 2002 02:45:35 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id CAA15213;
	Fri, 19 Apr 2002 02:27:08 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id CAA15186
	for <tsvwg@ns.ietf.org>; Fri, 19 Apr 2002 02:27:06 -0400 (EDT)
Received: from blaze.hcltech.com (pop.hclt-guindy.co.in [203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA11809
	for <tsvwg@ietf.org>; Fri, 19 Apr 2002 02:27:04 -0400 (EDT)
Received: from netlab.hcltech.com (arun.hclt-guindy.co.in [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id LAA04703;
	Fri, 19 Apr 2002 11:54:59 +0530
Message-ID: <3CBFB8A8.DF2AAF9@netlab.hcltech.com>
Date: Fri, 19 Apr 2002 11:56:48 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: "Buraga, Partha S" <p_buraga@trillium.com>
CC: "'tsvwg@ietf.org'" <tsvwg@ietf.org>, "'rrs@cisco.com'" <rrs@cisco.com>,
        "'Ivan.Arias-Rodriguez@nokia.com'" <Ivan.Arias-Rodriguez@nokia.com>,
        "'Andreas Jungmaier'" <ajung@exp-math.uni-essen.de>
References: <A1171C2850D8D4119D7B0002A508C76402A42A9A@bgsmsx91.iind.intel.com>
Content-Type: multipart/alternative;
 boundary="------------2B5C3B45D28C361353D58C14"
Subject: [Tsvwg] Re: HEARTBEAT failure
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


--------------2B5C3B45D28C361353D58C14
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Hi Partha,

"Buraga, Partha S" wrote:

> Hi
>
> In multihoming scenario,  a node has got three addresses x,y,z. In which,  x
> -- address ( NIC ) was failed.
>
> Application ( user of SCTP ) specifies the source address -- x,  for an
> outgoing HEARTBEAT chunk. Tho address -- x was failed, kernel deliver the
> packet to the destination thro interface y or z.
>

    Just a clarification here.....  The User of SCTP can specify only the Destination Address
to which Heartbeat is to be sent not the source address..... The source address (which is
filled in the IP header by SCTP... And IP in no way cares this setting in IP header, it sends
the Packet out through  the NIC card, which its routing table suggests...) will be filled by
SCTP before giving the packet to IP...

>
> But, when the HEARTBEAT - ACK responded by the peer, lands at interface --
> x, it will fail. And after ' n ' number of retries, Association will fail.
>

    Yes, this is definitely a problem.....  The algorithm we used to overcome this is....
I'll take the same example with three address x,y,z

Nbr_Addr = 3       (x,y,z)
Current source address ( Cur_Saddr) =  x        ie., the source address to be filled in IP_HDR
Maximum Retransmission Count (MAX_RTX_CNT) = 9
HB_RTX_THRSH =  MAX_RTX_CNT / Nbr_Addr
                              =  9 / 3 = 3


 Now whenever the Heartbeat timer expires (ie., when the RtxCounter is increased)
*RtxCntr ++
* Hb_RxtCntr++
* if (Hb_RtxCntr >= HB_RTX_THRSH)
      Cur_Saddr = NEXT_ADDR (Cur_Saddr)          ie., the next address which will be "y"
       Hb_RtxCntr = 0

    The logic is to try all the source address (from the pool we have) within the 'n' number of
retries....  This way, if the problem is with the source address by which we are sending, it
can be rectified.... The main purpose of sending HeartBeat, is to check whether the
remote interface is proper or not... So no harm changing the source address....

******* Another important point is that, if the implementation is not kernel based ..... then
some implementation of IP stack, can overwrite the Source Address filled by SCTP....
In that case... the above algo, doesnt have any effect...

Is the solution OK???


Thanks
-arun


>
> If this is the case, how can we establish association with the peer. I feel,
> this problem will araise when we specify source address to the outgoing
> chunk. If not, With IP_ANY, kernel can choose the address and association
> could be established.
>



>
> Can you please look into this and reply back.
>
> thanks in advance
> partha

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************


--------------2B5C3B45D28C361353D58C14
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
Hi Partha,
<p>"Buraga, Partha S" wrote:
<blockquote TYPE=CITE>Hi
<p>In multihoming scenario,&nbsp; a node has got three addresses x,y,z.
In which,&nbsp; x
<br>-- address ( NIC ) was failed.
<p>Application ( user of SCTP ) specifies the source address -- x,&nbsp;
for an
<br>outgoing HEARTBEAT chunk. Tho address -- x was failed, kernel deliver
the
<br>packet to the destination thro interface y or z.
<br>&nbsp;</blockquote>
&nbsp;&nbsp;&nbsp; Just a clarification here.....&nbsp; The User of SCTP
can specify only the Destination Address
<br>to which Heartbeat is to be sent not the source address..... The source
address (which is
<br>filled in the IP header by SCTP... And IP in no way cares this setting
in IP header, it sends
<br>the Packet out through&nbsp; the NIC card, which its routing table
suggests...) will be filled by
<br>SCTP before giving the packet to IP...
<blockquote TYPE=CITE>&nbsp;
<br>But, when the HEARTBEAT - ACK responded by the peer, lands at interface
--
<br>x, it will fail. And after ' n ' number of retries, Association will
fail.
<br>&nbsp;</blockquote>
&nbsp;&nbsp;&nbsp; Yes, this is definitely a problem.....&nbsp; The algorithm
we used to overcome this is....
<br>I'll take the same example with three address x,y,z
<p>Nbr_Addr = 3&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; (x,y,z)
<br>Current source address ( Cur_Saddr) =&nbsp; x&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
ie., the source address to be filled in IP_HDR
<br>Maximum Retransmission Count (MAX_RTX_CNT) = 9
<br>HB_RTX_THRSH =&nbsp; MAX_RTX_CNT / Nbr_Addr
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
=&nbsp; 9 / 3 = <b>3</b>
<br><b></b>&nbsp;<b></b>
<p><b>&nbsp;</b>Now whenever the Heartbeat timer expires (ie., when the
RtxCounter is increased)
<br>*RtxCntr ++
<br>* Hb_RxtCntr++
<br>* if (Hb_RtxCntr >= HB_RTX_THRSH)
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Cur_Saddr = NEXT_ADDR (Cur_Saddr)&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
ie., the next address which will be "y"
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Hb_RtxCntr = 0
<p>&nbsp;&nbsp;&nbsp; The logic is to try all the source address (from
the pool we have) within the 'n' number of
<br>retries....&nbsp; This way, if the problem is with the source address
by which we are sending, it
<br>can be rectified.... The main purpose of sending HeartBeat, is to check
whether the
<br>remote interface is proper or not... So no harm changing the source
address....
<p>******* Another important point is that, if the implementation is not
kernel based ..... then
<br>some implementation of IP stack, can overwrite the Source Address filled
by SCTP....
<br>In that case... the above algo, doesnt have any effect...
<p>Is the solution OK???
<br>&nbsp;
<p>Thanks
<br>-arun
<br>&nbsp;
<blockquote TYPE=CITE>&nbsp;
<br>If this is the case, how can we establish association with the peer.
I feel,
<br>this problem will araise when we specify source address to the outgoing
<br>chunk. If not, With IP_ANY, kernel can choose the address and association
<br>could be established.
<br>&nbsp;</blockquote>
&nbsp;
<blockquote TYPE=CITE>&nbsp;
<br>Can you please look into this and reply back.
<p>thanks in advance
<br>partha</blockquote>

<p>--
<br>****************************************************************
<br>V.Arun Prasad
<br>HCL Technologies Ltd.
<br>51, Jawaharlal Nehru Road,
<br>Ekkattuthangal,
<br>Guindy Industrial Estate,
<br>Chennai - 600097.
<p>Contact # : 9144 - 2334174
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
9144 - 2334181
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
extn : 233
<br>****************************************************************
<br>&nbsp;</html>

--------------2B5C3B45D28C361353D58C14--


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Apr 19 03:01:01 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA12176
	for <tsvwg-archive@odin.ietf.org>; Fri, 19 Apr 2002 03:01:00 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id DAA17531
	for tsvwg-archive@odin.ietf.org; Fri, 19 Apr 2002 03:01:01 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id CAA16711;
	Fri, 19 Apr 2002 02:47:18 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id CAA16670
	for <tsvwg@ns.ietf.org>; Fri, 19 Apr 2002 02:47:15 -0400 (EDT)
Received: from caduceus.fm.intel.com (fmr02.intel.com [192.55.52.25])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA12041
	for <tsvwg@ietf.org>; Fri, 19 Apr 2002 02:47:12 -0400 (EDT)
Received: from talaria.fm.intel.com (talaria.fm.intel.com [10.1.192.39])
	by caduceus.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.39 2002/04/15 17:47:23 root Exp $) with ESMTP id g3J6kUc07451
	for <tsvwg@ietf.org>; Fri, 19 Apr 2002 06:46:30 GMT
Received: from fmsmsxvs042.fm.intel.com (fmsmsxv042-1.fm.intel.com [132.233.48.110])
	by talaria.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.16 2002/04/15 17:47:00 root Exp $) with SMTP id g3J6n5w09325
	for <tsvwg@ietf.org>; Fri, 19 Apr 2002 06:49:05 GMT
Received: from fmsmsx29.FM.INTEL.COM ([132.233.42.29])
 by fmsmsxvs042.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002041823511604149
 ; Thu, 18 Apr 2002 23:51:16 -0700
Received: by fmsmsx29.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <2LKB9RAM>; Thu, 18 Apr 2002 23:47:12 -0700
Message-ID: <A1171C2850D8D4119D7B0002A508C76402A42A9E@bgsmsx91.iind.intel.com>
From: "Buraga, Partha S" <p_buraga@trillium.com>
To: "'Arun Prasad'" <arun@netlab.hcltech.com>,
        "Buraga, Partha S" <p_buraga@trillium.com>
Cc: "'tsvwg@ietf.org'" <tsvwg@ietf.org>, "'rrs@cisco.com'" <rrs@cisco.com>,
        "'Ivan.Arias-Rodriguez@nokia.com'" <Ivan.Arias-Rodriguez@nokia.com>,
        "'Andreas Jungmaier'" <ajung@exp-math.uni-essen.de>
Date: Thu, 18 Apr 2002 23:44:56 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01C1E76D.B7B309A0"
Subject: [Tsvwg] RE: HEARTBEAT failure
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01C1E76D.B7B309A0
Content-Type: text/plain;
	charset="ISO-8859-1"

Hi 
 
Here is my response inline.

-----Original Message-----
From: Arun Prasad [mailto:arun@netlab.hcltech.com]
Sent: Friday, April 19, 2002 11:57 AM
To: Buraga, Partha S
Cc: 'tsvwg@ietf.org'; 'rrs@cisco.com'; 'Ivan.Arias-Rodriguez@nokia.com';
'Andreas Jungmaier'
Subject: Re: HEARTBEAT failure


Hi Partha, 

"Buraga, Partha S" wrote: 


Hi 

In multihoming scenario,  a node has got three addresses x,y,z. In which,  x

-- address ( NIC ) was failed. 


Application ( user of SCTP ) specifies the source address -- x,  for an 
outgoing HEARTBEAT chunk. Tho address -- x was failed, kernel deliver the 
packet to the destination thro interface y or z. 
 

    Just a clarification here.....  The User of SCTP can specify only the
Destination Address 
to which Heartbeat is to be sent not the source address..... The source
address (which is 
filled in the IP header by SCTP... And IP in no way cares this setting in IP
header, it sends 
the Packet out through  the NIC card, which its routing table suggests...)
will be filled by 
SCTP before giving the packet to IP... 


But, when the HEARTBEAT - ACK responded by the peer, lands at interface -- 
x, it will fail. And after ' n ' number of retries, Association will fail. 
 

    Yes, this is definitely a problem.....  The algorithm we used to
overcome this is.... 
I'll take the same example with three address x,y,z 

Nbr_Addr = 3       (x,y,z) 
Current source address ( Cur_Saddr) =  x        ie., the source address to
be filled in IP_HDR 
Maximum Retransmission Count (MAX_RTX_CNT) = 9 
HB_RTX_THRSH =  MAX_RTX_CNT / Nbr_Addr 
                              =  9 / 3 = 3 
  


 Now whenever the Heartbeat timer expires (ie., when the RtxCounter is
increased) 
*RtxCntr ++ 
* Hb_RxtCntr++ 
* if (Hb_RtxCntr >= HB_RTX_THRSH) 
      Cur_Saddr = NEXT_ADDR (Cur_Saddr)          ie., the next address which
will be "y" 
       Hb_RtxCntr = 0 


    The logic is to try all the source address (from the pool we have)
within the 'n' number of 
retries....  This way, if the problem is with the source address by which we
are sending, it 
can be rectified.... The main purpose of sending HeartBeat, is to check
whether the 
remote interface is proper or not... So no harm changing the source
address.... 


******* Another important point is that, if the implementation is not kernel
based ..... then 
some implementation of IP stack, can overwrite the Source Address filled by
SCTP.... 
In that case... the above algo, doesnt have any effect... 


Is the solution OK???  


[Buraga, Partha S]  

I am working on User lever SCTP. Not kernel based.

One issue, with your reply that, you said as IP stack, can overwrite the
Source address filled by SCTP. But, this should not happen in IPV6 case.
Becas, drafts says, Application should have control on Source address and
Destination address of the packet. 

 [Buraga, Partha S] I have one Doubt, Will Kernel always selects a Source
address of the NIC, which is active ? 

thanks
partha 
  

Thanks 
-arun 
  



If this is the case, how can we establish association with the peer. I feel,

this problem will araise when we specify source address to the outgoing 
chunk. If not, With IP_ANY, kernel can choose the address and association 
could be established. 
 

  


Can you please look into this and reply back. 

thanks in advance 
partha

-- 
**************************************************************** 
V.Arun Prasad 
HCL Technologies Ltd. 
51, Jawaharlal Nehru Road, 
Ekkattuthangal, 
Guindy Industrial Estate, 
Chennai - 600097. 


Contact # : 9144 - 2334174 
            9144 - 2334181 
            extn : 233 
**************************************************************** 
  


------_=_NextPart_001_01C1E76D.B7B309A0
Content-Type: text/html;
	charset="ISO-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3DISO-8859-1">


<META content=3D"MSHTML 5.00.3502.4856" name=3DGENERATOR></HEAD>
<BODY>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN =
class=3D303364006-19042002>Hi=20
</SPAN></FONT></DIV>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN=20
class=3D303364006-19042002></SPAN></FONT>&nbsp;</DIV>
<DIV><FONT color=3D#0000ff face=3DArial size=3D2><SPAN =
class=3D303364006-19042002>Here=20
is my response inline.</SPAN></FONT></DIV>
<BLOCKQUOTE style=3D"MARGIN-RIGHT: 0px">
  <DIV align=3Dleft class=3DOutlookMessageHeader dir=3Dltr><FONT =
face=3DTahoma=20
  size=3D2>-----Original Message-----<BR><B>From:</B> Arun Prasad=20
  [mailto:arun@netlab.hcltech.com]<BR><B>Sent:</B> Friday, April 19, =
2002 11:57=20
  AM<BR><B>To:</B> Buraga, Partha S<BR><B>Cc:</B> 'tsvwg@ietf.org';=20
  'rrs@cisco.com'; 'Ivan.Arias-Rodriguez@nokia.com'; 'Andreas=20
  Jungmaier'<BR><B>Subject:</B> Re: HEARTBEAT =
failure<BR><BR></DIV></FONT>Hi=20
  Partha,=20
  <P>"Buraga, Partha S" wrote:=20
  <BLOCKQUOTE TYPE=3D"CITE">Hi=20
    <P>In multihoming scenario,&nbsp; a node has got three addresses =
x,y,z. In=20
    which,&nbsp; x <BR>-- address ( NIC ) was failed.=20
    <P>Application ( user of SCTP ) specifies the source address -- =
x,&nbsp; for=20
    an <BR>outgoing HEARTBEAT chunk. Tho address -- x was failed, =
kernel deliver=20
    the <BR>packet to the destination thro interface y or z.=20
  <BR>&nbsp;</P></BLOCKQUOTE>&nbsp;&nbsp;&nbsp; Just a clarification=20
  here.....&nbsp; The User of SCTP can specify only the Destination =
Address=20
  <BR>to which Heartbeat is to be sent not the source address..... The =
source=20
  address (which is <BR>filled in the IP header by SCTP... And IP in no =
way=20
  cares this setting in IP header, it sends <BR>the Packet out =
through&nbsp; the=20
  NIC card, which its routing table suggests...) will be filled by =
<BR>SCTP=20
  before giving the packet to IP...=20
  <BLOCKQUOTE TYPE=3D"CITE"> <BR>But, when the HEARTBEAT - ACK =
responded by the=20
    peer, lands at interface -- <BR>x, it will fail. And after ' n ' =
number of=20
    retries, Association will fail. =
<BR>&nbsp;</BLOCKQUOTE>&nbsp;&nbsp;&nbsp; Yes,=20
  this is definitely a problem.....&nbsp; The algorithm we used to =
overcome this=20
  is.... <BR>I'll take the same example with three address x,y,z=20
  <P>Nbr_Addr =3D 3&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; (x,y,z) =
<BR>Current source=20
  address ( Cur_Saddr) =3D&nbsp; =
x&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; ie.,=20
  the source address to be filled in IP_HDR <BR>Maximum Retransmission =
Count=20
  (MAX_RTX_CNT) =3D 9 <BR>HB_RTX_THRSH =3D&nbsp; MAX_RTX_CNT / Nbr_Addr =

  =
<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&n=
bsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&n=
bsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =3D&nbsp; 9 / 3 =3D <B>3</B> <BR><B></B>&nbsp;<B></B>=20
  <P><B>&nbsp;</B>Now whenever the Heartbeat timer expires (ie., when =
the=20
  RtxCounter is increased) <BR>*RtxCntr ++ <BR>* Hb_RxtCntr++ <BR>* if=20
  (Hb_RtxCntr &gt;=3D HB_RTX_THRSH) <BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; =
Cur_Saddr =3D=20
  NEXT_ADDR =
(Cur_Saddr)&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  ie., the next address which will be "y"=20
  <BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Hb_RtxCntr =3D 0=20
  <P>&nbsp;&nbsp;&nbsp; The logic is to try all the source address =
(from the=20
  pool we have) within the 'n' number of <BR>retries....&nbsp; This =
way, if the=20
  problem is with the source address by which we are sending, it =
<BR>can be=20
  rectified.... The main purpose of sending HeartBeat, is to check =
whether the=20
  <BR>remote interface is proper or not... So no harm changing the =
source=20
  address....=20
  <P>******* Another important point is that, if the implementation is =
not=20
  kernel based ..... then <BR>some implementation of IP stack, can =
overwrite the=20
  Source Address filled by SCTP.... <BR>In that case... the above algo, =
doesnt=20
  have any effect...=20
  <P>Is the solution OK???&nbsp;<FONT face=3DArial><FONT =
color=3D#0000ff><FONT=20
  size=3D2><SPAN class=3D303364006-19042002>&nbsp;</SPAN><BR><SPAN=20
  class=3D303364006-19042002></FONT></FONT></FONT><SPAN=20
  class=3D704153506-19042002><FONT color=3D#0000ff face=3DArial=20
  size=3D2></FONT></SPAN></SPAN></P>
  <P><SPAN class=3D303364006-19042002><SPAN =
class=3D704153506-19042002><FONT=20
  color=3D#0000ff face=3DArial size=3D2>[Buraga, Partha=20
  S]&nbsp;&nbsp;</FONT></SPAN><SPAN class=3D704153506-19042002><FONT =
color=3D#0000ff=20
  face=3DArial size=3D2></FONT></SPAN></SPAN></P>
  <P><SPAN class=3D303364006-19042002><SPAN =
class=3D704153506-19042002><FONT=20
  color=3D#0000ff face=3DArial size=3D2>I&nbsp;am working on User lever =
SCTP. Not=20
  kernel based.</FONT></SPAN></P>
  <P><FONT size=3D2><FONT color=3D#0000ff><FONT face=3DArial><SPAN=20
  class=3D704153506-19042002>One issue,&nbsp;with your reply that, you =
said as IP=20
  stack, can overwrite the Source address filled by SCTP. But, this =
should not=20
  happen in IPV6 case. Becas, drafts says, Application should have=20
  control&nbsp;on Source address and Destination address of the =
packet.<SPAN=20
  =
class=3D303364006-19042002>&nbsp;</SPAN></SPAN></FONT></FONT></FONT></P>=
<FONT=20
  face=3DArial><FONT color=3D#0000ff>
  <P><FONT size=3D2>&nbsp;<SPAN class=3D303364006-19042002>[Buraga, =
Partha S]&nbsp;I=20
  have one Doubt, Will Kernel always selects a Source address of the =
NIC, which=20
  is active ?&nbsp;</SPAN></FONT></P>
  <P></FONT></FONT></SPAN><SPAN class=3D303364006-19042002><FONT =
color=3D#0000ff=20
  face=3DArial size=3D2>thanks</FONT></SPAN><SPAN =
class=3D303364006-19042002><FONT=20
  color=3D#0000ff face=3DArial=20
  =
size=3D2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;=20
  partha</FONT>&nbsp;</SPAN><BR>&nbsp; </P>
  <P>Thanks <BR>-arun <BR>&nbsp;=20
  <BLOCKQUOTE TYPE=3D"CITE"> <BR>If this is the case, how can we =
establish=20
    association with the peer. I feel, <BR>this problem will araise =
when we=20
    specify source address to the outgoing <BR>chunk. If not, With =
IP_ANY,=20
    kernel can choose the address and association <BR>could be =
established.=20
  <BR>&nbsp;</BLOCKQUOTE>&nbsp;=20
  <BLOCKQUOTE TYPE=3D"CITE"> <BR>Can you please look into this and =
reply back.=20
    <P>thanks in advance <BR>partha</P></BLOCKQUOTE>
  <P>-- =
<BR>****************************************************************=20
  <BR>V.Arun Prasad <BR>HCL Technologies Ltd. <BR>51, Jawaharlal Nehru =
Road,=20
  <BR>Ekkattuthangal, <BR>Guindy Industrial Estate, <BR>Chennai - =
600097.=20
  <P>Contact # : 9144 - 2334174=20
  =
<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; =
9144 -=20
  2334181 =
<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  extn : 233=20
  <BR>****************************************************************=20
  <BR>&nbsp; </P></BLOCKQUOTE></BODY></HTML>

------_=_NextPart_001_01C1E76D.B7B309A0--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Apr 19 06:11:43 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA15045
	for <tsvwg-archive@odin.ietf.org>; Fri, 19 Apr 2002 06:11:43 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA29439
	for tsvwg-archive@odin.ietf.org; Fri, 19 Apr 2002 06:11:46 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA27468;
	Fri, 19 Apr 2002 05:53:52 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA27435
	for <tsvwg@ns.ietf.org>; Fri, 19 Apr 2002 05:53:49 -0400 (EDT)
Received: from irmgard.exp-math.uni-essen.de (irmgard.exp-math.uni-essen.de [132.252.150.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA14650;
	Fri, 19 Apr 2002 05:53:45 -0400 (EDT)
Received: from biene.exp-math.uni-essen.de (biene.exp-math.uni-essen.de [132.252.150.215])
	by irmgard.exp-math.uni-essen.de (8.12.2/8.12.2) with ESMTP id g3J9rksj052814;
	Fri, 19 Apr 2002 11:53:46 +0200
Content-Type: text/plain;
  charset="us-ascii"
From: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
Organization: University Essen
To: sigtran@ietf.org, tsvwg@ietf.org
Date: Fri, 19 Apr 2002 11:53:40 +0200
X-Mailer: KMail [version 1.4]
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Message-Id: <200204191153.42996.ajung@exp-math.uni-essen.de>
Content-Transfer-Encoding: 8bit
Subject: [Tsvwg] Virus alert...
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Dear members of these lists,

for quite a while  now I have been receiving emails with 
faked sender addresses, with the addresses naming (probably) subscribers
of these lists. 
These emails contain a (Windows) macro virus, an apparently empty
attachment, and some apparently private files (I got a number of
C source code files, which I probably shouldn't have got :-)

All of these mails seem to originate from a mail server at 
smtp.huawei.com with an IP address of 61.144.161.21

I have notified ABUSE, and some of the people whose name have been 
mis-used.

Should you have got such a mail containing MY name, please disregard
it, UNLESS it has the proper GnuPG signature - key available from my 
homepage.

I apalogize for having taken your time, but the issue seemed 
worthwhile to prevent this.

Regards,

Andreas Jungmaier

- -- 
- ------------------------------------------------------------------------
Dipl.-Ing. Andreas Jungmaier              |                         _
Computer Networking Technology Group      | ASCII ribbon campaign  ( )
University of Essen                       |  - against HTML email   X
http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8v+kkXAsLBNOCSsARAmjlAJ9Ho04QzgGH6cKG+lZ+SGauvRovIACfRhUY
gmBwfCpXBmUsgIW4vsBj+Vc=
=0VB8
-----END PGP SIGNATURE-----


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Apr 19 07:44:00 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA16537
	for <tsvwg-archive@odin.ietf.org>; Fri, 19 Apr 2002 07:44:00 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA03984
	for tsvwg-archive@odin.ietf.org; Fri, 19 Apr 2002 07:44:01 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA03013;
	Fri, 19 Apr 2002 07:27:46 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA02982
	for <tsvwg@ns.ietf.org>; Fri, 19 Apr 2002 07:27:42 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA16233
	for <tsvwg@ietf.org>; Fri, 19 Apr 2002 07:27:39 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3JBRUH33702;
	Fri, 19 Apr 2002 06:27:31 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CBFFF22.FDCC3985@stewart.chicago.il.us>
Date: Fri, 19 Apr 2002 06:27:30 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
CC: tsvwg@ietf.org, hn84456393@hn.hinet.net
Subject: Re: [Tsvwg] Virus alert...
References: <200204191153.42996.ajung@exp-math.uni-essen.de>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Andreas:

I think you are not alone :-)

I have been getting ton's of the same sorts of things ... not
from you but from others that I communicate with... for some
reason it is all on my cisco account... but since I don't
use windows it does not seem to bother me... except for the
annoyance that people may mistake me as sending them :-0

I would hope the folks in Tiawan that administor the
site will do something to stop all this.. The second cc 
address is supposed to be the administrator id...

Regards


R

Andreas Jungmaier wrote:
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Dear members of these lists,
> 
> for quite a while  now I have been receiving emails with
> faked sender addresses, with the addresses naming (probably) subscribers
> of these lists.
> These emails contain a (Windows) macro virus, an apparently empty
> attachment, and some apparently private files (I got a number of
> C source code files, which I probably shouldn't have got :-)
> 
> All of these mails seem to originate from a mail server at
> smtp.huawei.com with an IP address of 61.144.161.21
> 
> I have notified ABUSE, and some of the people whose name have been
> mis-used.
> 
> Should you have got such a mail containing MY name, please disregard
> it, UNLESS it has the proper GnuPG signature - key available from my
> homepage.
> 
> I apalogize for having taken your time, but the issue seemed
> worthwhile to prevent this.
> 
> Regards,
> 
> Andreas Jungmaier
> 
> - --
> - ------------------------------------------------------------------------
> Dipl.-Ing. Andreas Jungmaier              |                         _
> Computer Networking Technology Group      | ASCII ribbon campaign  ( )
> University of Essen                       |  - against HTML email   X
> http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.0.6 (GNU/Linux)
> Comment: For info see http://www.gnupg.org
> 
> iD8DBQE8v+kkXAsLBNOCSsARAmjlAJ9Ho04QzgGH6cKG+lZ+SGauvRovIACfRhUY
> gmBwfCpXBmUsgIW4vsBj+Vc=
> =0VB8
> -----END PGP SIGNATURE-----
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Apr 19 21:11:13 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA28803
	for <tsvwg-archive@odin.ietf.org>; Fri, 19 Apr 2002 21:11:13 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id VAA21185
	for tsvwg-archive@odin.ietf.org; Fri, 19 Apr 2002 21:11:15 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA20214;
	Fri, 19 Apr 2002 20:48:26 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA20166
	for <tsvwg@ns.ietf.org>; Fri, 19 Apr 2002 20:48:23 -0400 (EDT)
Received: from msgbas2.cos.agilent.com (msgbas2x.cos.agilent.com [192.25.240.37])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA26566;
	Fri, 19 Apr 2002 20:48:19 -0400 (EDT)
Received: from msgrel1.cos.agilent.com (msgrel1.cos.agilent.com [130.29.152.77])
	by msgbas2.cos.agilent.com (Postfix) with ESMTP
	id 4771E8B6; Fri, 19 Apr 2002 18:48:21 -0600 (MDT)
Received: from axcsbh2.cos.agilent.com (axcsbh2.cos.agilent.com [130.29.152.144])
	by msgrel1.cos.agilent.com (Postfix) with SMTP
	id C5F341A3; Fri, 19 Apr 2002 18:48:20 -0600 (MDT)
Received: from 130.29.152.144 by axcsbh2.cos.agilent.com (InterScan E-Mail VirusWall NT); Fri, 19 Apr 2002 18:48:20 -0600
Received: by axcsbh2.cos.agilent.com with Internet Mail Service (5.5.2653.19)
	id <262B91NG>; Fri, 19 Apr 2002 18:48:20 -0600
Message-ID: <1BEBA5E8600DD4119A50009027AF54A00B760A6A@axcs04.cos.agilent.com>
From: "THALER,PAT (A-Roseville,ex1)" <pat_thaler@agilent.com>
To: Andreas Jungmaier <ajung@exp-math.uni-essen.de>, sigtran@ietf.org,
        tsvwg@ietf.org
Subject: RE: [Tsvwg] Virus alert...
Date: Fri, 19 Apr 2002 18:48:18 -0600
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Andreas,

This sounds like the virus that I've encountered recently: w32.klez.h@mm or
one of the other w32.klez varients. The virus uses email addresses that it
finds on the infected system as the to and the from. If so, the IP address
of the source probably identifies an infected system. It also sends emails
that look like they are a bounce from the postmaster within your domain. 
See:
http://securityresponse.symantec.com/avcenter/venc/data/pf/w32.klez.h@mm.htm
l

for the details. The date on the virus definition there is April 17, 20002.

Pat



-----Original Message-----
From: Andreas Jungmaier [mailto:ajung@exp-math.uni-essen.de]
Sent: Friday, April 19, 2002 2:54 AM
To: sigtran@ietf.org; tsvwg@ietf.org
Subject: [Tsvwg] Virus alert...


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Dear members of these lists,

for quite a while  now I have been receiving emails with 
faked sender addresses, with the addresses naming (probably) subscribers
of these lists. 
These emails contain a (Windows) macro virus, an apparently empty
attachment, and some apparently private files (I got a number of
C source code files, which I probably shouldn't have got :-)

All of these mails seem to originate from a mail server at 
smtp.huawei.com with an IP address of 61.144.161.21

I have notified ABUSE, and some of the people whose name have been 
mis-used.

Should you have got such a mail containing MY name, please disregard
it, UNLESS it has the proper GnuPG signature - key available from my 
homepage.

I apalogize for having taken your time, but the issue seemed 
worthwhile to prevent this.

Regards,

Andreas Jungmaier

- -- 
- ------------------------------------------------------------------------
Dipl.-Ing. Andreas Jungmaier              |                         _
Computer Networking Technology Group      | ASCII ribbon campaign  ( )
University of Essen                       |  - against HTML email   X
http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8v+kkXAsLBNOCSsARAmjlAJ9Ho04QzgGH6cKG+lZ+SGauvRovIACfRhUY
gmBwfCpXBmUsgIW4vsBj+Vc=
=0VB8
-----END PGP SIGNATURE-----


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Apr 23 09:37:11 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA17167
	for <tsvwg-archive@odin.ietf.org>; Tue, 23 Apr 2002 09:37:11 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA00700
	for tsvwg-archive@odin.ietf.org; Tue, 23 Apr 2002 09:37:14 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA28719;
	Tue, 23 Apr 2002 09:10:22 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA28689
	for <tsvwg@ns.ietf.org>; Tue, 23 Apr 2002 09:10:18 -0400 (EDT)
Received: from caduceus.fm.intel.com (fmr02.intel.com [192.55.52.25])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA16152
	for <tsvwg@ietf.org>; Tue, 23 Apr 2002 09:10:13 -0400 (EDT)
Received: from petasus.fm.intel.com (petasus.fm.intel.com [10.1.192.37])
	by caduceus.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.39 2002/04/15 17:47:23 root Exp $) with ESMTP id g3ND9FR04389
	for <tsvwg@ietf.org>; Tue, 23 Apr 2002 13:09:16 GMT
Received: from fmsmsxvs042.fm.intel.com (fmsmsxv042-1.fm.intel.com [132.233.48.110])
	by petasus.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.16 2002/04/15 17:47:00 root Exp $) with SMTP id g3ND8S626491
	for <tsvwg@ietf.org>; Tue, 23 Apr 2002 13:08:28 GMT
Received: from fmsmsx26.fm.intel.com ([132.233.42.26])
 by fmsmsxvs042.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002042306141207158
 for <tsvwg@ietf.org>; Tue, 23 Apr 2002 06:14:12 -0700
Received: by fmsmsx26.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <2HL0NHAD>; Tue, 23 Apr 2002 06:10:02 -0700
Message-ID: <A1171C2850D8D4119D7B0002A508C764016983B4@bgsmsx91.iind.intel.com>
From: "Khan, Mohammad" <mo_khan@trillium.com>
To: "'tsvwg@ietf.org'" <tsvwg@ietf.org>
Cc: "Pillai, Hari" <h_pillai@trillium.com>
Date: Tue, 23 Apr 2002 06:07:48 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: [Tsvwg] End point and association Address conflict
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi All,
I have following querry regarding endpoint and association source addresses.

Suppose an Endpoint has three addresses X, Y and Z.
	An association which uses this endpoint and specifies only X and Y
addresses as source IP addresses. Is it allowed?


Regards
kashif


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Apr 23 10:54:03 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA20774
	for <tsvwg-archive@odin.ietf.org>; Tue, 23 Apr 2002 10:54:02 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA05626
	for tsvwg-archive@odin.ietf.org; Tue, 23 Apr 2002 10:54:06 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA04498;
	Tue, 23 Apr 2002 10:34:37 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA04467
	for <tsvwg@ns.ietf.org>; Tue, 23 Apr 2002 10:34:34 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-214.dallas.net [209.44.42.214])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA19711
	for <tsvwg@ietf.org>; Tue, 23 Apr 2002 10:34:28 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id JAA15364;
	Tue, 23 Apr 2002 09:34:20 -0500
Date: Tue, 23 Apr 2002 09:34:20 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: "Khan, Mohammad" <mo_khan@trillium.com>
Cc: "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] End point and association Address conflict
Message-ID: <20020423093420.B12323@openss7.org>
Reply-To: bidulock@openss7.org
References: <A1171C2850D8D4119D7B0002A508C764016983B4@bgsmsx91.iind.intel.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <A1171C2850D8D4119D7B0002A508C764016983B4@bgsmsx91.iind.intel.com>; from mo_khan@trillium.com on Tue, Apr 23, 2002 at 06:07:48AM -0700
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Khan,

Which transport protocol are you asking about?

--brian

On Tue, 23 Apr 2002, Khan, Mohammad wrote:

> Hi All,
> I have following querry regarding endpoint and association source addresses.
> 
> Suppose an Endpoint has three addresses X, Y and Z.
> 	An association which uses this endpoint and specifies only X and Y
> addresses as source IP addresses. Is it allowed?
> 
> 
> Regards
> kashif
> 
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Apr 23 15:11:44 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA02761
	for <tsvwg-archive@odin.ietf.org>; Tue, 23 Apr 2002 15:11:43 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA23507
	for tsvwg-archive@odin.ietf.org; Tue, 23 Apr 2002 15:11:47 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA21908;
	Tue, 23 Apr 2002 14:41:09 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA05115
	for <tsvwg@ns.ietf.org>; Tue, 23 Apr 2002 10:48:19 -0400 (EDT)
Received: from motgate2.mot.com (motgate2.mot.com [136.182.1.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA20333
	for <tsvwg@ietf.org>; Tue, 23 Apr 2002 10:48:14 -0400 (EDT)
Received: [from pobox3.mot.com (pobox3.mot.com [10.64.251.242]) by motgate2.mot.com (motgate2 2.1) with ESMTP id HAA22975; Tue, 23 Apr 2002 07:48:07 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox3.mot.com (MOT-pobox3 2.0) with ESMTP id HAA11078; Tue, 23 Apr 2002 07:35:30 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3NEm2j17531;
	Tue, 23 Apr 2002 09:48:02 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id JAA23466;
	Tue, 23 Apr 2002 09:48:09 -0500 (CDT)
Message-ID: <3CC574AD.1090303F@email.mot.com>
Date: Tue, 23 Apr 2002 09:50:21 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: "Khan, Mohammad" <mo_khan@trillium.com>
CC: tsvwg@ietf.org, "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] End point and association Address conflict
References: <A1171C2850D8D4119D7B0002A508C764016983B4@bgsmsx91.iind.intel.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi, Kashif,

Yes, you are allowed to do so, as long as you can guarantee that this
endpoint will never use Z as the source IP address in any packet sent
over this association. This is not hard to prevent if the endpoint has
full control over the source IP address selection routine.

-Qiaobing

"Khan, Mohammad" wrote:
> 
> Hi All,
> I have following querry regarding endpoint and association source addresses.
> 
> Suppose an Endpoint has three addresses X, Y and Z.
>         An association which uses this endpoint and specifies only X and Y
> addresses as source IP addresses. Is it allowed?
> 
> Regards
> kashif
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Apr 23 15:22:21 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA03236
	for <tsvwg-archive@odin.ietf.org>; Tue, 23 Apr 2002 15:22:21 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA23905
	for tsvwg-archive@odin.ietf.org; Tue, 23 Apr 2002 15:22:25 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA22482;
	Tue, 23 Apr 2002 14:55:02 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA22455
	for <tsvwg@ns.ietf.org>; Tue, 23 Apr 2002 14:55:00 -0400 (EDT)
Received: from hermes.fm.intel.com (fmr01.intel.com [192.55.52.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA02104
	for <tsvwg@ietf.org>; Tue, 23 Apr 2002 14:54:54 -0400 (EDT)
Received: from petasus.fm.intel.com (petasus.fm.intel.com [10.1.192.37])
	by hermes.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.41 2002/04/22 23:29:17 root Exp $) with ESMTP id g3NItw529921
	for <tsvwg@ietf.org>; Tue, 23 Apr 2002 18:55:58 GMT
Received: from fmsmsxvs042.fm.intel.com (fmsmsxv042-1.fm.intel.com [132.233.48.110])
	by petasus.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.16 2002/04/15 17:47:00 root Exp $) with SMTP id g3NIrM022205
	for <tsvwg@ietf.org>; Tue, 23 Apr 2002 18:53:22 GMT
Received: from FMSMSX018.fm.intel.com ([132.233.42.197])
 by fmsmsxvs042.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002042311590401624
 ; Tue, 23 Apr 2002 11:59:05 -0700
Received: by fmsmsx018.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <2WWS82NT>; Tue, 23 Apr 2002 11:54:53 -0700
Message-ID: <F1CE15E08172D4119247009027AE9D5019117392@FMSMSX37>
From: "Balani, Sandeep" <s_balani@trillium.com>
To: "'Qiaobing Xie'" <qxie1@email.mot.com>,
        "Khan, Mohammad" <mo_khan@trillium.com>
Cc: tsvwg@ietf.org
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Tue, 23 Apr 2002 11:54:46 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi Qiaobing, 
    I think the reason for asking such question was we have ADDIP & DELETEIP
procedure on an Association. 

Now Assume we have an Endpoint 

E1 =  [ IP-A, IP-B, and Port-X ] . 

Now lets say i have 2 associations on endpoint E1. 

A1 = [ Endpoint-E1 , DstAddr= D1, DstPort-P1 ] 
A2 = [ Endpoint-E2, DstAddr=D2,  DstPost-P2 ] 

Now when i do an  ADDIP [ for Address IP-C ] on association A1, actually i
am adding a source  IP address in Endpoint E1 i.e. [ IP-A, IP-B, IP-C,
Port-X ], so that I will find an Endpoint for an incoming packet. 

But IP-C is only used by Association A1, Not by A2.  That's the reason we
have an Association using Subset of IP addresses from an Endpoint.  

My Question is : Is it the intended behaviour in ADDIP/DELETEIP case ????

Many thanks
sandeep 
-----Original Message-----
From: Qiaobing Xie [mailto:qxie1@email.mot.com]
Sent: Tuesday, April 23, 2002 7:50 AM
To: Khan, Mohammad
Cc: tsvwg@ietf.org; Pillai, Hari
Subject: Re: [Tsvwg] End point and association Address conflict


Hi, Kashif,

Yes, you are allowed to do so, as long as you can guarantee that this
endpoint will never use Z as the source IP address in any packet sent
over this association. This is not hard to prevent if the endpoint has
full control over the source IP address selection routine.

-Qiaobing

"Khan, Mohammad" wrote:
> 
> Hi All,
> I have following querry regarding endpoint and association source
addresses.
> 
> Suppose an Endpoint has three addresses X, Y and Z.
>         An association which uses this endpoint and specifies only X and Y
> addresses as source IP addresses. Is it allowed?
> 
> Regards
> kashif
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 23 21:16:12 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA10915
	for <tsvwg-archive@odin.ietf.org>; Tue, 23 Apr 2002 21:16:11 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id VAA12571
	for tsvwg-archive@odin.ietf.org; Tue, 23 Apr 2002 21:16:16 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA11431;
	Tue, 23 Apr 2002 20:46:37 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA11391
	for <tsvwg@ns.ietf.org>; Tue, 23 Apr 2002 20:46:33 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA10607
	for <tsvwg@ietf.org>; Tue, 23 Apr 2002 20:46:27 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3O0kHH45577;
	Tue, 23 Apr 2002 19:46:17 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CC60059.3A33385A@stewart.chicago.il.us>
Date: Tue, 23 Apr 2002 19:46:17 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: "Balani, Sandeep" <s_balani@trillium.com>
CC: "'Qiaobing Xie'" <qxie1@email.mot.com>,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F1CE15E08172D4119247009027AE9D5019117392@FMSMSX37>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

The same scenario could exist outside of addip..

If I bind-all addresses.

setup an assocation (with IP-A/IP-B .. the ones I have)

now we hot-plug in a card.

I now setup a second association with all the
addresses I have (now IP-A/IP-B and IP-C).

You end up in the same place.. .and you must,
as Qiaobing said, assure that you only use IP addresses
that the peer knows about for sources..

Our ADDIP/DEL-IP allows this to happen as well.. we make
it so the user has to disable auto-asconf though to reach this state.


R


"Balani, Sandeep" wrote:
> 
> Hi Qiaobing,
>     I think the reason for asking such question was we have ADDIP & DELETEIP
> procedure on an Association.
> 
> Now Assume we have an Endpoint
> 
> E1 =  [ IP-A, IP-B, and Port-X ] .
> 
> Now lets say i have 2 associations on endpoint E1.
> 
> A1 = [ Endpoint-E1 , DstAddr= D1, DstPort-P1 ]
> A2 = [ Endpoint-E2, DstAddr=D2,  DstPost-P2 ]
> 
> Now when i do an  ADDIP [ for Address IP-C ] on association A1, actually i
> am adding a source  IP address in Endpoint E1 i.e. [ IP-A, IP-B, IP-C,
> Port-X ], so that I will find an Endpoint for an incoming packet.
> 
> But IP-C is only used by Association A1, Not by A2.  That's the reason we
> have an Association using Subset of IP addresses from an Endpoint.
> 
> My Question is : Is it the intended behaviour in ADDIP/DELETEIP case ????
> 
> Many thanks
> sandeep
> -----Original Message-----
> From: Qiaobing Xie [mailto:qxie1@email.mot.com]
> Sent: Tuesday, April 23, 2002 7:50 AM
> To: Khan, Mohammad
> Cc: tsvwg@ietf.org; Pillai, Hari
> Subject: Re: [Tsvwg] End point and association Address conflict
> 
> Hi, Kashif,
> 
> Yes, you are allowed to do so, as long as you can guarantee that this
> endpoint will never use Z as the source IP address in any packet sent
> over this association. This is not hard to prevent if the endpoint has
> full control over the source IP address selection routine.
> 
> -Qiaobing
> 
> "Khan, Mohammad" wrote:
> >
> > Hi All,
> > I have following querry regarding endpoint and association source
> addresses.
> >
> > Suppose an Endpoint has three addresses X, Y and Z.
> >         An association which uses this endpoint and specifies only X and Y
> > addresses as source IP addresses. Is it allowed?
> >
> > Regards
> > kashif
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 23 21:42:53 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA12220
	for <tsvwg-archive@odin.ietf.org>; Tue, 23 Apr 2002 21:42:53 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id VAA13850
	for tsvwg-archive@odin.ietf.org; Tue, 23 Apr 2002 21:42:55 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id VAA12675;
	Tue, 23 Apr 2002 21:19:26 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id VAA12645
	for <tsvwg@optimus.ietf.org>; Tue, 23 Apr 2002 21:19:23 -0400 (EDT)
Received: from hermes.fm.intel.com (fmr01.intel.com [192.55.52.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA10964
	for <tsvwg@ietf.org>; Tue, 23 Apr 2002 21:19:10 -0400 (EDT)
Received: from talaria.fm.intel.com (talaria.fm.intel.com [10.1.192.39])
	by hermes.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.41 2002/04/22 23:29:17 root Exp $) with ESMTP id g3O1JSc04729
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 01:19:28 GMT
Received: from fmsmsxvs042.fm.intel.com (fmsmsxv042-1.fm.intel.com [132.233.48.110])
	by talaria.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.16 2002/04/15 17:47:00 root Exp $) with SMTP id g3O1KVW17378
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 01:20:31 GMT
Received: from FMSMSX018.fm.intel.com ([132.233.42.197])
 by fmsmsxvs042.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002042318223612472
 ; Tue, 23 Apr 2002 18:22:36 -0700
Received: by fmsmsx018.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <2WWS0BF3>; Tue, 23 Apr 2002 18:18:25 -0700
Message-ID: <F1CE15E08172D4119247009027AE9D5019117396@FMSMSX37>
From: "Balani, Sandeep" <s_balani@trillium.com>
To: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        "Balani, Sandeep" <s_balani@trillium.com>
Cc: "'Qiaobing Xie'" <qxie1@email.mot.com>,
        "Khan, Mohammad"<mo_khan@trillium.com>, tsvwg@ietf.org
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Tue, 23 Apr 2002 18:18:21 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


Thanks Randy & Qiaobing, 
  Although i would like to add something here.  

So the inference of this discussion is : The source IP Addresses in an
association can be subset of Endpoint addresses.   Is this statment correct
? 

In the above case an Implmentation "MUST" make sure that whatever source IP
addresses are communicated during association intialisation should be used
while sending outoing packets. ( I.e. control U'r Src IP Addresses )

Also to extend the same logic, In the example I have shown below If i
terminate Association A1, I should also delete the entry of IP-C address in
Endpoint E1 Correct ????? 


Lastly this may never be TRUE but I would like to ask this question anyways.
In the example below I did'nt communicated IP-C on Association A2, But Lets
take a wild case that somehow I receive a SCTP packet on [ IP-C, Port-X,
D2, Port-P2] then SCTP will accept this Packet Is'nt it ?  As u will find
endpoint valid & Association  also .

Thanks
Sandeep





-----Original Message-----
From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
Sent: Tuesday, April 23, 2002 5:46 PM
To: Balani, Sandeep
Cc: 'Qiaobing Xie'; Khan, Mohammad; tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict


The same scenario could exist outside of addip..

If I bind-all addresses.

setup an assocation (with IP-A/IP-B .. the ones I have)

now we hot-plug in a card.

I now setup a second association with all the
addresses I have (now IP-A/IP-B and IP-C).

You end up in the same place.. .and you must,
as Qiaobing said, assure that you only use IP addresses
that the peer knows about for sources..

Our ADDIP/DEL-IP allows this to happen as well.. we make
it so the user has to disable auto-asconf though to reach this state.


R


"Balani, Sandeep" wrote:
> 
> Hi Qiaobing,
>     I think the reason for asking such question was we have ADDIP &
DELETEIP
> procedure on an Association.
> 
> Now Assume we have an Endpoint
> 
> E1 =  [ IP-A, IP-B, and Port-X ] .
> 
> Now lets say i have 2 associations on endpoint E1.
> 
> A1 = [ Endpoint-E1 , DstAddr= D1, DstPort-P1 ]
> A2 = [ Endpoint-E2, DstAddr=D2,  DstPost-P2 ]
> 
> Now when i do an  ADDIP [ for Address IP-C ] on association A1, actually i
> am adding a source  IP address in Endpoint E1 i.e. [ IP-A, IP-B, IP-C,
> Port-X ], so that I will find an Endpoint for an incoming packet.
> 
> But IP-C is only used by Association A1, Not by A2.  That's the reason we
> have an Association using Subset of IP addresses from an Endpoint.
> 
> My Question is : Is it the intended behaviour in ADDIP/DELETEIP case ????
> 
> Many thanks
> sandeep
> -----Original Message-----
> From: Qiaobing Xie [mailto:qxie1@email.mot.com]
> Sent: Tuesday, April 23, 2002 7:50 AM
> To: Khan, Mohammad
> Cc: tsvwg@ietf.org; Pillai, Hari
> Subject: Re: [Tsvwg] End point and association Address conflict
> 
> Hi, Kashif,
> 
> Yes, you are allowed to do so, as long as you can guarantee that this
> endpoint will never use Z as the source IP address in any packet sent
> over this association. This is not hard to prevent if the endpoint has
> full control over the source IP address selection routine.
> 
> -Qiaobing
> 
> "Khan, Mohammad" wrote:
> >
> > Hi All,
> > I have following querry regarding endpoint and association source
> addresses.
> >
> > Suppose an Endpoint has three addresses X, Y and Z.
> >         An association which uses this endpoint and specifies only X and
Y
> > addresses as source IP addresses. Is it allowed?
> >
> > Regards
> > kashif
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 23 23:14:37 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA14417
	for <tsvwg-archive@odin.ietf.org>; Tue, 23 Apr 2002 23:14:37 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id XAA17698
	for tsvwg-archive@odin.ietf.org; Tue, 23 Apr 2002 23:14:40 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id WAA16628;
	Tue, 23 Apr 2002 22:45:27 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id WAA16595
	for <tsvwg@optimus.ietf.org>; Tue, 23 Apr 2002 22:45:24 -0400 (EDT)
Received: from india.mercurykr.com ([165.133.13.60])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA14111
	for <tsvwg@ietf.org>; Tue, 23 Apr 2002 22:44:48 -0400 (EDT)
Received: from vijay (Archana [165.133.13.38])
	by india.mercurykr.com (8.8.8+Sun/8.8.8) with SMTP id IAA00810
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 08:05:52 -0600 (GMT)
MIME-Version: 1.0
Message-Id: <3CC61C93.000003.89639@vijay.india.mercurykr.com>
Date: Wed, 24 Apr 2002 08:16:43 +0530 (India Standard Time)
Content-Type: Text/Plain
X-Mailer: IncrediMail 2001 (1500245)
From: "Vijay Kumar Choudhary" <vijay@mercurykr.com>
X-Priority: 3
X-FID: FLAVOR00-NONE-0000-0000-000000000000
To: <tsvwg@ietf.org>
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id WAA16596
Subject: [Tsvwg] sctp process!!!!
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Hi All,

This is general ques............Just i want to know.....Can the
single process of SCTP handle all the things (i mean sending data and
receiving data)......or Do we need to have two separate process one for
sending data and another for receiving data.

How many processes are there in the kernel implementaion of TCP??

Regards,
Vijay

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 23 23:51:56 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA15274
	for <tsvwg-archive@odin.ietf.org>; Tue, 23 Apr 2002 23:51:56 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id XAA19043
	for tsvwg-archive@odin.ietf.org; Tue, 23 Apr 2002 23:51:59 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id XAA18109;
	Tue, 23 Apr 2002 23:30:35 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id XAA18084
	for <tsvwg@optimus.ietf.org>; Tue, 23 Apr 2002 23:30:32 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-214.dallas.net [209.44.42.214])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA14636
	for <tsvwg@ietf.org>; Tue, 23 Apr 2002 23:30:27 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id WAA25267;
	Tue, 23 Apr 2002 22:30:01 -0500
Date: Tue, 23 Apr 2002 22:30:01 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Vijay Kumar Choudhary <vijay@mercurykr.com>
Cc: tsvwg@ietf.org
Subject: Re: [Tsvwg] sctp process!!!!
Message-ID: <20020423223001.A23395@openss7.org>
Reply-To: bidulock@openss7.org
References: <3CC61C93.000003.89639@vijay.india.mercurykr.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <3CC61C93.000003.89639@vijay.india.mercurykr.com>; from vijay@mercurykr.com on Wed, Apr 24, 2002 at 08:16:43AM +0530
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Vijay,

On Wed, 24 Apr 2002, Vijay Kumar Choudhary wrote:

> Hi All,
> 
> This is general ques............Just i want to know.....Can the
> single process of SCTP handle all the things (i mean sending data and
> receiving data)......or Do we need to have two separate process one for
> sending data and another for receiving data.

One is enough.

> 
> How many processes are there in the kernel implementaion of TCP??

One.

--brian

> 
> Regards,
> Vijay
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr 24 00:11:52 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA15677
	for <tsvwg-archive@odin.ietf.org>; Wed, 24 Apr 2002 00:11:52 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id AAA20332
	for tsvwg-archive@odin.ietf.org; Wed, 24 Apr 2002 00:11:53 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id XAA19130;
	Tue, 23 Apr 2002 23:54:34 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id XAA19101
	for <tsvwg@optimus.ietf.org>; Tue, 23 Apr 2002 23:54:31 -0400 (EDT)
Received: from hermes.fm.intel.com (fmr01.intel.com [192.55.52.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA15302
	for <tsvwg@ietf.org>; Tue, 23 Apr 2002 23:54:24 -0400 (EDT)
Received: from talaria.fm.intel.com (talaria.fm.intel.com [10.1.192.39])
	by hermes.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.41 2002/04/22 23:29:17 root Exp $) with ESMTP id g3O3tTP25177
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 03:55:29 GMT
Received: from fmsmsxvs042.fm.intel.com (fmsmsxvs042.fm.intel.com [132.233.42.128])
	by talaria.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.16 2002/04/15 17:47:00 root Exp $) with SMTP id g3O3uXM12033
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 03:56:33 GMT
Received: from fmsmsx26.fm.intel.com ([132.233.42.26])
 by fmsmsxvs042.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002042320583928606
 for <tsvwg@ietf.org>; Tue, 23 Apr 2002 20:58:39 -0700
Received: by fmsmsx26.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <2HL0N0RR>; Tue, 23 Apr 2002 20:54:28 -0700
Message-ID: <A1171C2850D8D4119D7B0002A508C764016983B6@bgsmsx91.iind.intel.com>
From: "Khan, Mohammad" <mo_khan@trillium.com>
To: "'tsvwg@ietf.org'" <tsvwg@ietf.org>
Cc: "Pillai, Hari" <h_pillai@trillium.com>
Date: Tue, 23 Apr 2002 20:51:43 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: [Tsvwg] Source IP address Control
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi All,
One concern i have regarding SCTP source address control.
Can we keep the IP header source address of an outgoing sctp packet as one
which is down and this is known to peer through heartbeats
1)Should peer discard this data it receives from down interface address.
2)Should peer check the state of the peer address it is sending SACK,
HEARTBEAT-ACK, INIT-ACK and COOKIE-ACK.

Motive of doing this is implementation of Dynamic IP address modification
when i don't know the status of local interface IP address and one node e.g.
has four interface IP address A, B, C and D. One Endpoint E1 is using IP
addresses A , B, C and D and has opened an association A1 which has source
IP addresses A and B. If  I open IP_ANY socket for all the addresses on this
host and SCTP sends some data to the peer on association A1with source
address as A but even if A is down IP kernal shall send this packet out
through some other interface keeping header IP source address unchanged to
the network and the peer would receive the data but peer is knowing that
this interface A is down. Is it legal.

Regards
kashif khan

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr 24 01:19:11 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA16570
	for <tsvwg-archive@odin.ietf.org>; Wed, 24 Apr 2002 01:19:11 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id BAA02297
	for tsvwg-archive@odin.ietf.org; Wed, 24 Apr 2002 01:19:11 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id AAA21709;
	Wed, 24 Apr 2002 00:49:25 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id AAA21680
	for <tsvwg@optimus.ietf.org>; Wed, 24 Apr 2002 00:49:22 -0400 (EDT)
Received: from auemail2.firewall.lucent.com (auemail2.lucent.com [192.11.223.163])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA16292
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 00:49:21 -0400 (EDT)
Received: from horh1.emsr.lucent.com (h135-17-1-40.lucent.com [135.17.1.40])
	by auemail2.firewall.lucent.com (Switch-2.1.3/Switch-2.1.0) with ESMTP id g3O4mjZ23759;
	Wed, 24 Apr 2002 00:48:45 -0400 (EDT)
Received: from new-wopr.eng.ascend.com by horh1.emsr.lucent.com (8.9.3+Sun/EMS-1.5 Solaris/emsr)
	id AAA22525 for ; Wed, 24 Apr 2002 00:48:44 -0400 (EDT)
Received: from grigri.eng.ascend.com (grigri.eng.ascend.com [135.140.53.45])
	by new-wopr.eng.ascend.com (8.10.2+Sun/8.10.2) with ESMTP id g3O4mir25583;
	Tue, 23 Apr 2002 21:48:44 -0700 (PDT)
Received: from lucent.com by grigri.eng.ascend.com (8.8.8+Sun/SMI-SVR4)
	id VAA05668; Tue, 23 Apr 2002 21:48:43 -0700 (PDT)
Message-ID: <3CC639B5.820AC1D3@lucent.com>
Date: Tue, 23 Apr 2002 21:51:01 -0700
From: Jacob Heitz <jheitz@lucent.com>
Organization: Lucent Technologies, Internetworking Systems, Software Engineering, 
 Alameda
X-Mailer: Mozilla 4.77 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: "Khan, Mohammad" <mo_khan@trillium.com>
CC: "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] Source IP address Control
References: <A1171C2850D8D4119D7B0002A508C764016983B6@bgsmsx91.iind.intel.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

I don't think it is illegal, but why would you want
to do this ?

If your peer receives a data chunk with source address
A, he will, with high probability, send a SACK back to
address A. He may SACK to another address, for example,
if he has his own DATA to send and piggybacks a SACK
on it. You should accept a SACK to any address belonging
to your endpoint. The main point is that you cannot
expect your peer to know that A is down and therefore expect
him to send the SACK to B. He MAY do this, but you cannot
expect that he will.

"Khan, Mohammad" wrote:
> 
> Hi All,
> One concern i have regarding SCTP source address control.
> Can we keep the IP header source address of an outgoing sctp packet as one
> which is down and this is known to peer through heartbeats
> 1)Should peer discard this data it receives from down interface address.
> 2)Should peer check the state of the peer address it is sending SACK,
> HEARTBEAT-ACK, INIT-ACK and COOKIE-ACK.
> 
> Motive of doing this is implementation of Dynamic IP address modification
> when i don't know the status of local interface IP address and one node e.g.
> has four interface IP address A, B, C and D. One Endpoint E1 is using IP
> addresses A , B, C and D and has opened an association A1 which has source
> IP addresses A and B. If  I open IP_ANY socket for all the addresses on this
> host and SCTP sends some data to the peer on association A1with source
> address as A but even if A is down IP kernal shall send this packet out
> through some other interface keeping header IP source address unchanged to
> the network and the peer would receive the data but peer is knowing that
> this interface A is down. Is it legal.
> 
> Regards
> kashif khan
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
// Jacob Heitz. Tel:510-747-2917

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr 24 01:31:02 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA16678
	for <tsvwg-archive@odin.ietf.org>; Wed, 24 Apr 2002 01:31:02 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id BAA04117
	for tsvwg-archive@odin.ietf.org; Wed, 24 Apr 2002 01:31:03 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id AAA21909;
	Wed, 24 Apr 2002 00:58:42 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id AAA21882
	for <tsvwg@optimus.ietf.org>; Wed, 24 Apr 2002 00:58:40 -0400 (EDT)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA16371
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 00:58:37 -0400 (EDT)
Received: from netlab.hcltech.com (arun.hclt-guindy.co.in [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id KAA03475;
	Wed, 24 Apr 2002 10:25:56 +0530
Message-ID: <3CC63B48.7A8BA76A@netlab.hcltech.com>
Date: Wed, 24 Apr 2002 10:27:44 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: "Balani, Sandeep" <s_balani@trillium.com>
CC: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        "'Qiaobing Xie'" <qxie1@email.mot.com>,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F1CE15E08172D4119247009027AE9D5019117396@FMSMSX37>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

hi sandeep,

"Balani, Sandeep" wrote:

> Thanks Randy & Qiaobing,
>   Although i would like to add something here.
>
> So the inference of this discussion is : The source IP Addresses in an
> association can be subset of Endpoint addresses.   Is this statment correct
> ?
>
> In the above case an Implmentation "MUST" make sure that whatever source IP
> addresses are communicated during association intialisation should be used
> while sending outoing packets. ( I.e. control U'r Src IP Addresses )
>

    Yes thats true, but its not always that easy to implement ie., if we implement
a generalised SCTP stack, covering many OS... then its really tough to keep track
of this feature (full control of Src IP Addresses) in the OS....
for example:
    Windows 2000 supports IP_HDR_INCL option ( ie., adding our own Source
Ip Addresses) in its raw socket implementation, but it doesnt guarentee that the
Src address filled by the user will be going in the IP Packet sent from that machine...
In this case the subset concept will fail.....
But for Linux, the OS guarentees the source Ip Address given by the user will be
filled in the IP packet sent.....
So Its more of the OS dependent... (and the OS can change this implementation
any time....) ... The thing is, we should have a generalised API for Raw sockets, as
we have for other protocols (like TCP, UDP etc...)

>
> Also to extend the same logic, In the example I have shown below If i
> terminate Association A1, I should also delete the entry of IP-C address in
> Endpoint E1 Correct ?????
>



>
> Lastly this may never be TRUE but I would like to ask this question anyways.
> In the example below I did'nt communicated IP-C on Association A2, But Lets
> take a wild case that somehow I receive a SCTP packet on [ IP-C, Port-X,
> D2, Port-P2] then SCTP will accept this Packet Is'nt it ?  As u will find
> endpoint valid & Association  also .
>

    The Receiving endpoint will search the Association for which the packet has
come with the following parameters { SrcAddr, SrcPort, RemAddr, RemPort}
In the above case since the SrcAddr is a different one (that is not in the Association's
pool) the received packet will not hit on the Association A2...
So no probs....


Thanks
-arun

>
> Thanks
> Sandeep
>
> -----Original Message-----
> From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: Tuesday, April 23, 2002 5:46 PM
> To: Balani, Sandeep
> Cc: 'Qiaobing Xie'; Khan, Mohammad; tsvwg@ietf.org
> Subject: Re: [Tsvwg] End point and association Address conflict
>
> The same scenario could exist outside of addip..
>
> If I bind-all addresses.
>
> setup an assocation (with IP-A/IP-B .. the ones I have)
>
> now we hot-plug in a card.
>
> I now setup a second association with all the
> addresses I have (now IP-A/IP-B and IP-C).
>
> You end up in the same place.. .and you must,
> as Qiaobing said, assure that you only use IP addresses
> that the peer knows about for sources..
>
> Our ADDIP/DEL-IP allows this to happen as well.. we make
> it so the user has to disable auto-asconf though to reach this state.
>
> R
>
> "Balani, Sandeep" wrote:
> >
> > Hi Qiaobing,
> >     I think the reason for asking such question was we have ADDIP &
> DELETEIP
> > procedure on an Association.
> >
> > Now Assume we have an Endpoint
> >
> > E1 =  [ IP-A, IP-B, and Port-X ] .
> >
> > Now lets say i have 2 associations on endpoint E1.
> >
> > A1 = [ Endpoint-E1 , DstAddr= D1, DstPort-P1 ]
> > A2 = [ Endpoint-E2, DstAddr=D2,  DstPost-P2 ]
> >
> > Now when i do an  ADDIP [ for Address IP-C ] on association A1, actually i
> > am adding a source  IP address in Endpoint E1 i.e. [ IP-A, IP-B, IP-C,
> > Port-X ], so that I will find an Endpoint for an incoming packet.
> >
> > But IP-C is only used by Association A1, Not by A2.  That's the reason we
> > have an Association using Subset of IP addresses from an Endpoint.
> >
> > My Question is : Is it the intended behaviour in ADDIP/DELETEIP case ????
> >
> > Many thanks
> > sandeep
> > -----Original Message-----
> > From: Qiaobing Xie [mailto:qxie1@email.mot.com]
> > Sent: Tuesday, April 23, 2002 7:50 AM
> > To: Khan, Mohammad
> > Cc: tsvwg@ietf.org; Pillai, Hari
> > Subject: Re: [Tsvwg] End point and association Address conflict
> >
> > Hi, Kashif,
> >
> > Yes, you are allowed to do so, as long as you can guarantee that this
> > endpoint will never use Z as the source IP address in any packet sent
> > over this association. This is not hard to prevent if the endpoint has
> > full control over the source IP address selection routine.
> >
> > -Qiaobing
> >
> > "Khan, Mohammad" wrote:
> > >
> > > Hi All,
> > > I have following querry regarding endpoint and association source
> > addresses.
> > >
> > > Suppose an Endpoint has three addresses X, Y and Z.
> > >         An association which uses this endpoint and specifies only X and
> Y
> > > addresses as source IP addresses. Is it allowed?
> > >
> > > Regards
> > > kashif
> > >
> > > _______________________________________________
> > > tsvwg mailing list
> > > tsvwg@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/tsvwg
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
>
> --
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr 24 07:36:11 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA29884
	for <tsvwg-archive@odin.ietf.org>; Wed, 24 Apr 2002 07:36:11 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA22696
	for tsvwg-archive@odin.ietf.org; Wed, 24 Apr 2002 07:36:13 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA20376;
	Wed, 24 Apr 2002 07:10:27 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA20349
	for <tsvwg@optimus.ietf.org>; Wed, 24 Apr 2002 07:10:24 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA29365
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 07:10:21 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3OB9xH46666;
	Wed, 24 Apr 2002 06:10:00 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CC69287.DBD308B7@stewart.chicago.il.us>
Date: Wed, 24 Apr 2002 06:09:59 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: "Balani, Sandeep" <s_balani@trillium.com>
CC: "'Qiaobing Xie'" <qxie1@email.mot.com>,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F1CE15E08172D4119247009027AE9D5019117396@FMSMSX37>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

"Balani, Sandeep" wrote:
> 
> Thanks Randy & Qiaobing,
>   Although i would like to add something here.
> 
> So the inference of this discussion is : The source IP Addresses in an
> association can be subset of Endpoint addresses.   Is this statment correct
> ?
> 
> In the above case an Implmentation "MUST" make sure that whatever source IP
> addresses are communicated during association intialisation should be used
> while sending outoing packets. ( I.e. control U'r Src IP Addresses )
> 
Yes. If you don't use ALL source addresses then you must
tightly control what you place in your outbound IP packets :>

This CAN causes issues with some O/S's.. for those you may
only be able to allow the bind-all type symantic..

> Also to extend the same logic, In the example I have shown below If i
> terminate Association A1, I should also delete the entry of IP-C address in
> Endpoint E1 Correct ?????

I am a bit confused by the above. Does E1 have IP-C or not? 

I thought your example was that E1 bound all 3 A/B/C but for
some strange reason did NOT send C on one but did on another.

For our implementation an ENDPOINT binds a SET of addresses or
ALL addresses. So the address IP-C would NOT leave the
endpoint once the association ended...

> 
> Lastly this may never be TRUE but I would like to ask this question anyways.
> In the example below I did'nt communicated IP-C on Association A2, But Lets
> take a wild case that somehow I receive a SCTP packet on [ IP-C, Port-X,
> D2, Port-P2] then SCTP will accept this Packet Is'nt it ?  As u will find
> endpoint valid & Association  also .
> 
Well, I am not sure how this could even happen.. the peer only knows
A/B so it is really a non-issue. But I will say the answer is
implementation dependant. Some implementations MAY not find the
association... Others MAY.. I think the point is moot since the
situation is not possible...

R



> 
> -----Original Message-----
> From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: Tuesday, April 23, 2002 5:46 PM
> To: Balani, Sandeep
> Cc: 'Qiaobing Xie'; Khan, Mohammad; tsvwg@ietf.org
> Subject: Re: [Tsvwg] End point and association Address conflict
> 
> The same scenario could exist outside of addip..
> 
> If I bind-all addresses.
> 
> setup an assocation (with IP-A/IP-B .. the ones I have)
> 
> now we hot-plug in a card.
> 
> I now setup a second association with all the
> addresses I have (now IP-A/IP-B and IP-C).
> 
> You end up in the same place.. .and you must,
> as Qiaobing said, assure that you only use IP addresses
> that the peer knows about for sources..
> 
> Our ADDIP/DEL-IP allows this to happen as well.. we make
> it so the user has to disable auto-asconf though to reach this state.
> 
> R
> 
> "Balani, Sandeep" wrote:
> >
> > Hi Qiaobing,
> >     I think the reason for asking such question was we have ADDIP &
> DELETEIP
> > procedure on an Association.
> >
> > Now Assume we have an Endpoint
> >
> > E1 =  [ IP-A, IP-B, and Port-X ] .
> >
> > Now lets say i have 2 associations on endpoint E1.
> >
> > A1 = [ Endpoint-E1 , DstAddr= D1, DstPort-P1 ]
> > A2 = [ Endpoint-E2, DstAddr=D2,  DstPost-P2 ]
> >
> > Now when i do an  ADDIP [ for Address IP-C ] on association A1, actually i
> > am adding a source  IP address in Endpoint E1 i.e. [ IP-A, IP-B, IP-C,
> > Port-X ], so that I will find an Endpoint for an incoming packet.
> >
> > But IP-C is only used by Association A1, Not by A2.  That's the reason we
> > have an Association using Subset of IP addresses from an Endpoint.
> >
> > My Question is : Is it the intended behaviour in ADDIP/DELETEIP case ????
> >
> > Many thanks
> > sandeep
> > -----Original Message-----
> > From: Qiaobing Xie [mailto:qxie1@email.mot.com]
> > Sent: Tuesday, April 23, 2002 7:50 AM
> > To: Khan, Mohammad
> > Cc: tsvwg@ietf.org; Pillai, Hari
> > Subject: Re: [Tsvwg] End point and association Address conflict
> >
> > Hi, Kashif,
> >
> > Yes, you are allowed to do so, as long as you can guarantee that this
> > endpoint will never use Z as the source IP address in any packet sent
> > over this association. This is not hard to prevent if the endpoint has
> > full control over the source IP address selection routine.
> >
> > -Qiaobing
> >
> > "Khan, Mohammad" wrote:
> > >
> > > Hi All,
> > > I have following querry regarding endpoint and association source
> > addresses.
> > >
> > > Suppose an Endpoint has three addresses X, Y and Z.
> > >         An association which uses this endpoint and specifies only X and
> Y
> > > addresses as source IP addresses. Is it allowed?
> > >
> > > Regards
> > > kashif
> > >
> > > _______________________________________________
> > > tsvwg mailing list
> > > tsvwg@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/tsvwg
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> --
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr 24 07:37:29 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA29908
	for <tsvwg-archive@odin.ietf.org>; Wed, 24 Apr 2002 07:37:28 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA22714
	for tsvwg-archive@odin.ietf.org; Wed, 24 Apr 2002 07:37:30 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA20515;
	Wed, 24 Apr 2002 07:11:48 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA20419
	for <tsvwg@optimus.ietf.org>; Wed, 24 Apr 2002 07:11:42 -0400 (EDT)
Received: from caduceus.fm.intel.com (fmr02.intel.com [192.55.52.25])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA29389
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 07:11:40 -0400 (EDT)
Received: from petasus.fm.intel.com (petasus.fm.intel.com [10.1.192.37])
	by caduceus.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.41 2002/04/22 23:29:17 root Exp $) with ESMTP id g3OBArv03177
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 11:10:53 GMT
Received: from fmsmsxvs043.fm.intel.com (fmsmsxv043-1.fm.intel.com [132.233.48.128])
	by petasus.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.16 2002/04/15 17:47:00 root Exp $) with SMTP id g3OBA5d08920
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 11:10:05 GMT
Received: from fmsmsx019.fm.intel.com ([132.233.42.130])
 by fmsmsxvs043.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002042404104409510
 ; Wed, 24 Apr 2002 04:10:44 -0700
Received: by fmsmsx019.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <2XMHC0W0>; Wed, 24 Apr 2002 04:11:40 -0700
Message-ID: <A1171C2850D8D4119D7B0002A508C764016983BF@bgsmsx91.iind.intel.com>
From: "Khan, Mohammad" <mo_khan@trillium.com>
To: "'Jacob Heitz'" <jheitz@lucent.com>,
        "Khan, Mohammad" <mo_khan@trillium.com>
Cc: "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: RE: [Tsvwg] Source IP address Control
Date: Wed, 24 Apr 2002 04:09:21 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi Jacob,
As u said "you cannot expect your peer to know that A is down and therefore
expect
him to send the SACK to B". Through heartbeats peer is knowing that A is
down and what extra it has to do is to check the status of the address on
which it is sending SACK if it is down it has to send the SACK to another
address of the same association which is alive. As far as piggybacking the
data with sack is concerned offcourse sending SCTP shall send it to the
locally know alive address of the peer.  
Regards
kashif

-----Original Message-----
From: Jacob Heitz [mailto:jheitz@lucent.com]
Sent: Wednesday, April 24, 2002 10:21 AM
To: Khan, Mohammad
Cc: 'tsvwg@ietf.org'; Pillai, Hari
Subject: Re: [Tsvwg] Source IP address Control


I don't think it is illegal, but why would you want
to do this ?

If your peer receives a data chunk with source address
A, he will, with high probability, send a SACK back to
address A. He may SACK to another address, for example,
if he has his own DATA to send and piggybacks a SACK
on it. You should accept a SACK to any address belonging
to your endpoint. The main point is that you cannot
expect your peer to know that A is down and therefore expect
him to send the SACK to B. He MAY do this, but you cannot
expect that he will.

"Khan, Mohammad" wrote:
> 
> Hi All,
> One concern i have regarding SCTP source address control.
> Can we keep the IP header source address of an outgoing sctp packet as one
> which is down and this is known to peer through heartbeats
> 1)Should peer discard this data it receives from down interface address.
> 2)Should peer check the state of the peer address it is sending SACK,
> HEARTBEAT-ACK, INIT-ACK and COOKIE-ACK.
> 
> Motive of doing this is implementation of Dynamic IP address modification
> when i don't know the status of local interface IP address and one node
e.g.
> has four interface IP address A, B, C and D. One Endpoint E1 is using IP
> addresses A , B, C and D and has opened an association A1 which has source
> IP addresses A and B. If  I open IP_ANY socket for all the addresses on
this
> host and SCTP sends some data to the peer on association A1with source
> address as A but even if A is down IP kernal shall send this packet out
> through some other interface keeping header IP source address unchanged to
> the network and the peer would receive the data but peer is knowing that
> this interface A is down. Is it legal.
> 
> Regards
> kashif khan
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
// Jacob Heitz. Tel:510-747-2917

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr 24 08:59:20 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA02307
	for <tsvwg-archive@odin.ietf.org>; Wed, 24 Apr 2002 08:59:20 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA27565
	for tsvwg-archive@odin.ietf.org; Wed, 24 Apr 2002 08:59:22 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA26045;
	Wed, 24 Apr 2002 08:34:23 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA26018
	for <tsvwg@optimus.ietf.org>; Wed, 24 Apr 2002 08:34:20 -0400 (EDT)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA01559
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 08:34:16 -0400 (EDT)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir04nok.ntc.nokia.com (esvir04nokt.ntc.nokia.com [172.21.143.36])
	by mgw-x2.nokia.com (Switch-2.2.1/Switch-2.2.0) with ESMTP id g3OCYVF00682
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 15:34:32 +0300 (EET DST)
Received: from esebh004.NOE.Nokia.com (unverified) by esvir04nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5a75070c14ac158f24078@esvir04nok.ntc.nokia.com>;
 Wed, 24 Apr 2002 15:34:12 +0300
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh004.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Wed, 24 Apr 2002 15:34:12 +0300
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Wed, 24 Apr 2002 15:34:12 +0300
Message-ID: <58AEDE27D48F884285007ABF827FF26308EFA8@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] End point and association Address conflict
Thread-Index: AcHrgLPPX3mCRu2mQm2YCFuNLg6zOAABEmMQ
To: <randall@stewart.chicago.il.us>, <s_balani@trillium.com>
Cc: <qxie1@email.mot.com>, <mo_khan@trillium.com>, <tsvwg@ietf.org>
X-OriginalArrivalTime: 24 Apr 2002 12:34:12.0622 (UTC) FILETIME=[56BC62E0:01C1EB8C]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id IAA26019
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi all!

	I would like to make some comments, just to be sure that I understand the logic of the dynamic addition/deletion of IP addresses for SCTP in the right way... As the subject is a little bit confusing, I'll try to be as clear as I can... which means a long mail... :-(

> "Balani, Sandeep" wrote:
> > 
> > Thanks Randy & Qiaobing,
> >   Although i would like to add something here.
> > 
> > So the inference of this discussion is : The source IP 
> Addresses in an
> > association can be subset of Endpoint addresses.   Is this 
> statment correct
> > ?

	In my understanding, this is the whole picture. Your machine has several different NICs, let's say each of them with a single IP address. Imagine that those IP addresses are IPa, IPb, IPc and IPd... You can use all of them or just a subset of them, plus a port number, to create an endpoint. As always, if two endpoints use the same port, they cannot share the use of any address (i.e., that address cannot be used in more than one endpoint). Endpoints E1(IPa, IPb; Port 1), E2(IPc, IPd; Port 1), and E3(IPa, IPb, IPc; Port 2) could coexist without problems...

	Let's say that you have created E1. Then you are listening on IPa and IPb with Port number 1. When you receive SCTP datagrams (from unknown peer addresses) to any of these two addresses using Port 1, you should answer in the right way depending the type of chunks it carries (INIT, COOKIE ECHO, SHUTDOWN ACK, OOTB...).

	However, when you create an association, it initially uses ALL the IP addresses of that endpoint (no matter if your user started it or if the peer was the one sending the INIT chunk)... This is my understanding... :-/ If you create the association in a different endpoint, you can use a different set of addresses, though...

> > In the above case an Implmentation "MUST" make sure that 
> whatever source IP
> > addresses are communicated during association intialisation 
> should be used
> > while sending outoing packets. ( I.e. control U'r Src IP Addresses )
> > 
> Yes. If you don't use ALL source addresses then you must
> tightly control what you place in your outbound IP packets :>
> 
> This CAN causes issues with some O/S's.. for those you may
> only be able to allow the bind-all type symantic..

	You must only use the addresses you use... I mean, as a (hopefully) easy to understand example, the same happens with the port number. If your association uses port number 23471, then, you must use that port, and not 23472 (even though another association might be using it)... Addresses/port is what determines each side of the association, so you should not use any Address or Port that is not part of the association...

> > Also to extend the same logic, In the example I have shown 
> below If i
> > terminate Association A1, I should also delete the entry of 
> IP-C address in
> > Endpoint E1 Correct ?????

	This is a somehow difficult thing to think about... This is precisely what made me change almost all the structure of my implementation...

	I would like people to comment on my comment, to see if I'm right...

	You can have several endpoints. Let's say we have the three defined above.

	Let's start for the easy thing. If you have an association created in E3, A1E3, if you erase any of the source addresses of that association, you DON'T have to erase that address from E3. I.e., if you erase IPc, then, you would still process SCTP datagrams directed to IPc and Port 2 (as directed to E3 and coming from an unknown peer), HOWEVER, you won't consider that datagram as being directed to A1E3 even if the peer of A1E3 sent that datagram (if it contains DATA, it will be an OOTB datagram).

	But this brings some other problems... Imagine that the peer of association A1E3 is IPz, Port 65535. A1E3 (after erasing IPc) will in fact be described by: [(IPa, IPb; Port 2), (IPz; Port 65535)]. If now we receive a datagram from IPz/65535, directed to IPc/2, what happens? If it doesn't carry an INIT or COOKIE ECHO, the answer is easy, just discard it, send back an ABORT or SHUTDOWN COMPLETE depending on the chunks inside... But, if the peer (IPz; Port 65535) is in fact starting a new association (sending INIT or COOKIE ECHO), what should we do? We cannot open a new association, let's say A2E3, containing all the source addresses of E3, because then we wouldn't know, for example, if a chunk from IPz/65535 and directed to IPa/2 goes to A1E3 or A2E3...

	In my case, I don't allow this. Before establishing a new association I check if this kind of problem happens...

	Now, let's say that you have another association, A1E1, created in E1, and using both IPa and IPb with Port 1, whose peer is (IPy; Port 65535). If you have another association, A1E2, created in E2, using IPc and IPd with Port 1, whose peer is the same as A1E1's one (IPy; Port 65535), what happens if A1E1 want's to add IPc as another source address?

	In my case I check again, that this wouldn't conflict with any existing association before I allow the user to do so...

	Imagine that A1E2 does not exist. In that case you would be able to add IPc to A1E1. All the datagrams directed to IPc/2 will be considered as directed to E2 if THERE ISN'T any already established association using IPc/2 whose peer uses the source address/port of the incoming datagram... I.e., if there isn't any other association established but A1E2 (with the added IPc), all the datagrams directed to IPc/2 will be treated as from an unknown peer and directed to E2, except from the ones coming from IPy/65535, that are part of A1E2...

	After adding IPc to A1E1, the user will NOT be able to create a new association in E2 whose peer is IPy/65535, for the same reasons stated above...

	So, you never change the endpoint, just the association... This creates some problems, though... That's why AddIP extension is in my opinion a somehow difficult extension to add to your implementation...

> I am a bit confused by the above. Does E1 have IP-C or not? 
> 
> I thought your example was that E1 bound all 3 A/B/C but for
> some strange reason did NOT send C on one but did on another.
> 
> For our implementation an ENDPOINT binds a SET of addresses or
> ALL addresses. So the address IP-C would NOT leave the
> endpoint once the association ended...

	Yes, but an association created within an endpoint uses always ALL the source addresses of that endpoint, right?

> > Lastly this may never be TRUE but I would like to ask this 
> question anyways.
> > In the example below I did'nt communicated IP-C on 
> Association A2, But Lets
> > take a wild case that somehow I receive a SCTP packet on [ 
> IP-C, Port-X,
> > D2, Port-P2] then SCTP will accept this Packet Is'nt it ?  
> As u will find
> > endpoint valid & Association  also .
> > 
> Well, I am not sure how this could even happen.. the peer only knows
> A/B so it is really a non-issue. But I will say the answer is
> implementation dependant. Some implementations MAY not find the
> association... Others MAY.. I think the point is moot since the
> situation is not possible...

	I'm with Randall here... If you have not sent yet the ASCONF, then, the peer should not send anything (unless you are connected to a future teller server ;->)... However, as the source address is considered in use (only when receiving datagrams, not when sending them) even before the ASCONF ACK is received, I guess that an implementation could do it either way... In my opinion it would be more reasonable not to accept that datagram as part of the association, though... :-/

	That's all!

	BR Iván Arias Rodríguez

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr 24 10:59:27 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA17450
	for <tsvwg-archive@odin.ietf.org>; Wed, 24 Apr 2002 10:59:27 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA06362
	for tsvwg-archive@odin.ietf.org; Wed, 24 Apr 2002 10:59:30 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA03777;
	Wed, 24 Apr 2002 10:29:42 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA03742
	for <tsvwg@optimus.ietf.org>; Wed, 24 Apr 2002 10:29:39 -0400 (EDT)
Received: from hoemail1.firewall.lucent.com (hoemail1.lucent.com [192.11.226.161])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA09740
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 10:29:35 -0400 (EDT)
Received: from horh1.emsr.lucent.com (h135-17-1-40.lucent.com [135.17.1.40])
	by hoemail1.firewall.lucent.com (Switch-2.1.3/Switch-2.1.0) with ESMTP id g3OESv720605;
	Wed, 24 Apr 2002 10:28:58 -0400 (EDT)
Received: from new-wopr.eng.ascend.com by horh1.emsr.lucent.com (8.9.3+Sun/EMS-1.5 Solaris/emsr)
	id KAA13819 for ; Wed, 24 Apr 2002 10:28:56 -0400 (EDT)
Received: from grigri.eng.ascend.com (grigri.eng.ascend.com [135.140.53.45])
	by new-wopr.eng.ascend.com (8.10.2+Sun/8.10.2) with ESMTP id g3OESur22308;
	Wed, 24 Apr 2002 07:28:56 -0700 (PDT)
Received: from lucent.com by grigri.eng.ascend.com (8.8.8+Sun/SMI-SVR4)
	id HAA14997; Wed, 24 Apr 2002 07:28:55 -0700 (PDT)
Message-ID: <3CC6C1B2.D09236A2@lucent.com>
Date: Wed, 24 Apr 2002 07:31:14 -0700
From: Jacob Heitz <jheitz@lucent.com>
Organization: Lucent Technologies, Internetworking Systems, Software Engineering, 
 Alameda
X-Mailer: Mozilla 4.77 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: "Khan, Mohammad" <mo_khan@trillium.com>
CC: "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] Source IP address Control
References: <A1171C2850D8D4119D7B0002A508C764016983BF@bgsmsx91.iind.intel.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Are you trying to add a requirement to SCTP like this:

  If an endpoint knows that a transport address of a peer
  endpoint is down, then it MUST send SACK to a different
  address if one is available.

?

Best of luck to you.

"Khan, Mohammad" wrote:
> 
> Hi Jacob,
> As u said "you cannot expect your peer to know that A is down and therefore
> expect
> him to send the SACK to B". Through heartbeats peer is knowing that A is
> down and what extra it has to do is to check the status of the address on
> which it is sending SACK if it is down it has to send the SACK to another
> address of the same association which is alive. As far as piggybacking the
> data with sack is concerned offcourse sending SCTP shall send it to the
> locally know alive address of the peer.
> Regards
> kashif
> 
> -----Original Message-----
> From: Jacob Heitz [mailto:jheitz@lucent.com]
> Sent: Wednesday, April 24, 2002 10:21 AM
> To: Khan, Mohammad
> Cc: 'tsvwg@ietf.org'; Pillai, Hari
> Subject: Re: [Tsvwg] Source IP address Control
> 
> I don't think it is illegal, but why would you want
> to do this ?
> 
> If your peer receives a data chunk with source address
> A, he will, with high probability, send a SACK back to
> address A. He may SACK to another address, for example,
> if he has his own DATA to send and piggybacks a SACK
> on it. You should accept a SACK to any address belonging
> to your endpoint. The main point is that you cannot
> expect your peer to know that A is down and therefore expect
> him to send the SACK to B. He MAY do this, but you cannot
> expect that he will.
> 
> "Khan, Mohammad" wrote:
> >
> > Hi All,
> > One concern i have regarding SCTP source address control.
> > Can we keep the IP header source address of an outgoing sctp packet as one
> > which is down and this is known to peer through heartbeats
> > 1)Should peer discard this data it receives from down interface address.
> > 2)Should peer check the state of the peer address it is sending SACK,
> > HEARTBEAT-ACK, INIT-ACK and COOKIE-ACK.
> >
> > Motive of doing this is implementation of Dynamic IP address modification
> > when i don't know the status of local interface IP address and one node
> e.g.
> > has four interface IP address A, B, C and D. One Endpoint E1 is using IP
> > addresses A , B, C and D and has opened an association A1 which has source
> > IP addresses A and B. If  I open IP_ANY socket for all the addresses on
> this
> > host and SCTP sends some data to the peer on association A1with source
> > address as A but even if A is down IP kernal shall send this packet out
> > through some other interface keeping header IP source address unchanged to
> > the network and the peer would receive the data but peer is knowing that
> > this interface A is down. Is it legal.
> >
> > Regards
> > kashif khan

-- 
// Jacob Heitz. Tel:510-747-2917

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Apr 24 12:38:07 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA25409
	for <tsvwg-archive@odin.ietf.org>; Wed, 24 Apr 2002 12:38:07 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id MAA15044
	for tsvwg-archive@odin.ietf.org; Wed, 24 Apr 2002 12:38:09 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA13428;
	Wed, 24 Apr 2002 12:22:42 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA13385
	for <tsvwg@optimus.ietf.org>; Wed, 24 Apr 2002 12:22:38 -0400 (EDT)
Received: from mail.eecis.udel.edu (louie.udel.edu [128.4.40.12])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id MAA21333
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 12:22:33 -0400 (EDT)
Received: from ren.eecis.udel.edu by mail.eecis.udel.edu id aa15725;
          24 Apr 2002 12:22 EDT
Date: Wed, 24 Apr 2002 12:22:07 -0400 (EDT)
From: "Armando L. Caro Jr." <acaro@mail.eecis.udel.edu>
To: "Khan, Mohammad" <mo_khan@trillium.com>
cc: "'Jacob Heitz'" <jheitz@lucent.com>, "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: RE: [Tsvwg] Source IP address Control
In-Reply-To: <A1171C2850D8D4119D7B0002A508C764016983BF@bgsmsx91.iind.intel.com>
Message-ID: <Pine.GSO.4.33.0204241215490.7361-100000@ren.eecis.udel.edu>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi Mohammad,

I haven't followed this thread carefully. So, if I have missed something
let me know.

I think there is a flaw in your argument. If A is down, the peer may
notice that HBs are not being acked, but the peer may not realize yet that
A is down. So until Path.Max.Retrans is reached, the peer still
considers A to be up. Hence, the peer may very well send the SACK to A.

Now, if Path.Max.Retrans has been reached, then the peer DOES knows that A
is down. However, as Jacob pointed out, you are attempting to add an
additional requirement to the protocol by requiring that the peer send the
SACK to B.

~armando

0--                                                --0
| Armando L. Caro Jr.     |       acaro@cis.udel.edu |
| University of Delaware  |  www.cis.udel.edu/~acaro |
0--                                                --0

On Wed, 24 Apr 2002, Khan, Mohammad wrote:

> Hi Jacob,
> As u said "you cannot expect your peer to know that A is down and therefore
> expect
> him to send the SACK to B". Through heartbeats peer is knowing that A is
> down and what extra it has to do is to check the status of the address on
> which it is sending SACK if it is down it has to send the SACK to another
> address of the same association which is alive. As far as piggybacking the
> data with sack is concerned offcourse sending SCTP shall send it to the
> locally know alive address of the peer.
> Regards
> kashif



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Wed Apr 24 16:58:52 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA07420
	for <tsvwg-archive@odin.ietf.org>; Wed, 24 Apr 2002 16:58:52 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA00280
	for tsvwg-archive@odin.ietf.org; Wed, 24 Apr 2002 16:58:52 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA29492;
	Wed, 24 Apr 2002 16:38:18 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA29464
	for <tsvwg@ns.ietf.org>; Wed, 24 Apr 2002 16:38:14 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA06394
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 16:38:05 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3OKbKH47602;
	Wed, 24 Apr 2002 15:37:22 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CC7176A.DAA9D0A0@stewart.chicago.il.us>
Date: Wed, 24 Apr 2002 15:36:58 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: s_balani@trillium.com, qxie1@email.mot.com, mo_khan@trillium.com,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFA8@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Ivan.Arias-Rodriguez@nokia.com wrote:
> 
>         Hi all!
> 
>         I would like to make some comments, just to be sure that I understand the logic of the dynamic addition/deletion of IP addresses for SCTP in the right way... As the subject is a little bit confusing, I'll try to be as clear as I can... which means a long mail... :-(
> 
> > "Balani, Sandeep" wrote:
> > >
> > > Thanks Randy & Qiaobing,
> > >   Although i would like to add something here.
> > >
> > > So the inference of this discussion is : The source IP
> > Addresses in an
> > > association can be subset of Endpoint addresses.   Is this
> > statment correct
> > > ?
> 
>         In my understanding, this is the whole picture. Your machine has several different NICs, let's say each of them with a single IP address. Imagine that those IP addresses are IPa, IPb, IPc and IPd... You can use all of them or just a subset of them, plus a port number, to create an endpoint. As always, if two endpoints use the same port, they cannot share the use of any address (i.e., that address cannot be used in more than one endpoint). Endpoints E1(IPa, IPb; Port 1), E2(IPc, IPd; Port 1), and E3(IPa, IPb, IPc; Port 2) could coexist without problems...
> 
>         Let's say that you have created E1. Then you are listening on IPa and IPb with Port number 1. When you receive SCTP datagrams (from unknown peer addresses) to any of these two addresses using Port 1, you should answer in the right way depending the type of chunks it carries (INIT, COOKIE ECHO, SHUTDOWN ACK, OOTB...).
> 
>         However, when you create an association, it initially uses ALL the IP addresses of that endpoint (no matter if your user started it or if the peer was the one sending the INIT chunk)... This is my understanding... :-/ If you create the association in a different endpoint, you can use a different set of addresses, though...

I agree 100%.. when you send out your INIT or INIT-ACK you MUST USE ALL
of your addresses that are BOUND to the endpoint.. You should NEVER
send a "sub-set" of the endpoints addresses.. 

> 
> > > In the above case an Implmentation "MUST" make sure that
> > whatever source IP
> > > addresses are communicated during association intialisation
> > should be used
> > > while sending outoing packets. ( I.e. control U'r Src IP Addresses )
> > >
> > Yes. If you don't use ALL source addresses then you must
> > tightly control what you place in your outbound IP packets :>
> >
> > This CAN causes issues with some O/S's.. for those you may
> > only be able to allow the bind-all type symantic..
> 
>         You must only use the addresses you use... I mean, as a (hopefully) easy to understand example, the same happens with the port number. If your association uses port number 23471, then, you must use that port, and not 23472 (even though another association might be using it)... Addresses/port is what determines each side of the association, so you should not use any Address or Port that is not part of the association...
> 

Exactly!

> > > Also to extend the same logic, In the example I have shown
> > below If i
> > > terminate Association A1, I should also delete the entry of
> > IP-C address in
> > > Endpoint E1 Correct ?????
> 
>         This is a somehow difficult thing to think about... This is precisely what made me change almost all the structure of my implementation...
> 
>         I would like people to comment on my comment, to see if I'm right...
> 
>         You can have several endpoints. Let's say we have the three defined above.
> 
>         Let's start for the easy thing. If you have an association created in E3, A1E3, if you erase any of the source addresses of that association, you DON'T have to erase that address from E3. I.e., if you erase IPc, then, you would still process SCTP datagrams directed to IPc and Port 2 (as directed to E3 and coming from an unknown peer), HOWEVER, you won't consider that datagram as being directed to A1E3 even if the peer of A1E3 sent that datagram (if it contains DATA, it will be an OOTB datagram).
> 
>         But this brings some other problems... Imagine that the peer of association A1E3 is IPz, Port 65535. A1E3 (after erasing IPc) will in fact be described by: [(IPa, IPb; Port 2), (IPz; Port 65535)]. If now we receive a datagram from IPz/65535, directed to IPc/2, what happens? If it doesn't carry an INIT or COOKIE ECHO, the answer is easy, just discard it, send back an ABORT or SHUTDOWN COMPLETE depending on the chunks inside... But, if the peer (IPz; Port 65535) is in fact starting a new association (sending INIT or COOKIE ECHO), what should we do? We cannot open a new association, let's say A2E3, containing all the source addresses of E3, because then we wouldn't know, for example, if a chunk from IPz/65535 and directed to IPa/2 goes to A1E3 or A2E3...
> 
>         In my case, I don't allow this. Before establishing a new association I check if this kind of problem happens...
> 
I too would not allow this. I must also say that if this corner case
were to occur I would think it a bit unusual.. it is a bit odd... this
is why I think it is more than safe to NOT allow this..

>         Now, let's say that you have another association, A1E1, created in E1, and using both IPa and IPb with Port 1, whose peer is (IPy; Port 65535). If you have another association, A1E2, created in E2, using IPc and IPd with Port 1, whose peer is the same as A1E1's one (IPy; Port 65535), what happens if A1E1 want's to add IPc as another source address?
> 

But how could this even be legal... if 

E1={IPA, IPB + Port 1}
and
E2={IPC, IPD + Port 1}

Trying to put IPC in A1E1 means you are attempting to move
IPC from an already existing endpoint aka E2. You can't it is
a binding conflict... So you should never reach the state of
trying this since you have the IP address (IPC) already in
another endpoint E2...


>         In my case I check again, that this wouldn't conflict with any existing association before I allow the user to do so...

No, I think it is a conflict of an existing Endpoint NOT association.

> 
>         Imagine that A1E2 does not exist. In that case you would be able to add IPc to A1E1. 

No you can't do that. IPc belongs to E2 NOT E1.. so you can NOT add
a IP address to an association that does not have THAT IP address
BOUND to the ENDPOINT!

>         All the datagrams directed to IPc/2 will be considered as directed to E2 if THERE ISN'T any already established association using IPc/2 whose peer uses the source address/port of the incoming datagram... I.e., if there isn't any other association established but A1E2 (with the added IPc), all the datagrams directed to IPc/2 will be treated as from an unknown peer and directed to E2, except from the ones coming from IPy/65535, that are part of A1E2...

> 
>         After adding IPc to A1E1, the user will NOT be able to create a new association in E2 whose peer is IPy/65535, for the same reasons stated above...
> 
No if E2 exists and has bound IPC/IPB your SHOULD NOT ALLOW IPC to be
added to A1E1... E1 does NOT have IPC bound.

The basic concept is that a ENDPOINT must BIND the address FIRST
BEFORE you can use the dynamic IP address add. If the IP address
exists in another ENDPOINT then it is the same as if you try
to bind an address that someone else has... i.e. you get an error.

In the case of binding all ... there is no problem. you always
get all IP addresses on the machine... But once you start subset
binding the ENDPOINT must bind the address BEFORE you can add it
to ANY of the ENDPOINTS associations...


>         So, you never change the endpoint, just the association... This creates some problems, though... That's why AddIP extension is in my opinion a somehow difficult extension to add to your implementation...
> 
I think that depends. If one has a well structured implementation it
is no problem... You must keep the rules that you can't add an
IP address unless it is BOUND to the endpoint.. This is not
stated in the draft since the draft does not go in to the details
of the internal API's of the socket interface (for example). But 
you MUST structure your endpoint -to-> associations releationships
properly...



> > I am a bit confused by the above. Does E1 have IP-C or not?
> >
> > I thought your example was that E1 bound all 3 A/B/C but for
> > some strange reason did NOT send C on one but did on another.
> >
> > For our implementation an ENDPOINT binds a SET of addresses or
> > ALL addresses. So the address IP-C would NOT leave the
> > endpoint once the association ended...
> 
>         Yes, but an association created within an endpoint uses always ALL the source addresses of that endpoint, right?
> 
> > > Lastly this may never be TRUE but I would like to ask this
> > question anyways.
> > > In the example below I did'nt communicated IP-C on
> > Association A2, But Lets
> > > take a wild case that somehow I receive a SCTP packet on [
> > IP-C, Port-X,
> > > D2, Port-P2] then SCTP will accept this Packet Is'nt it ?
> > As u will find
> > > endpoint valid & Association  also .
> > >
> > Well, I am not sure how this could even happen.. the peer only knows
> > A/B so it is really a non-issue. But I will say the answer is
> > implementation dependant. Some implementations MAY not find the
> > association... Others MAY.. I think the point is moot since the
> > situation is not possible...
> 
>         I'm with Randall here... If you have not sent yet the ASCONF, then, the peer should not send anything (unless you are connected to a future teller server ;->)... However, as the source address is considered in use (only when receiving datagrams, not when sending them) even before the ASCONF ACK is received, I guess that an implementation could do it either way... In my opinion it would be more reasonable not to accept that datagram as part of the association, though... :-/
> 
If one does not while your ASCONF is "in flight" and you have not
recieved
the ASCONF-ACK.. one could always DROP the DG... the peer will
retransmit
it.. My implementation does not do this though.. it gets put into
a "special" status as I accept from it but wont source from it
until the ASCONF-ACK arrives...

I think either way is ok though...

R

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Wed Apr 24 17:06:16 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA08245
	for <tsvwg-archive@odin.ietf.org>; Wed, 24 Apr 2002 17:06:15 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA01402
	for tsvwg-archive@odin.ietf.org; Wed, 24 Apr 2002 17:06:07 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA29778;
	Wed, 24 Apr 2002 16:46:39 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA29748
	for <tsvwg@ns.ietf.org>; Wed, 24 Apr 2002 16:46:36 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA06596
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 16:46:30 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3OKhSH47621;
	Wed, 24 Apr 2002 15:43:29 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CC718F0.C5D186FA@stewart.chicago.il.us>
Date: Wed, 24 Apr 2002 15:43:28 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: "Armando L. Caro Jr." <acaro@mail.eecis.udel.edu>
CC: "Khan, Mohammad" <mo_khan@trillium.com>,
        "'Jacob Heitz'" <jheitz@lucent.com>,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] Source IP address Control
References: <Pine.GSO.4.33.0204241215490.7361-100000@ren.eecis.udel.edu>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Armando:

I don't see this as a needed requirement... it is just an implementation
decision on how they wish to handle SACK's.

I know one of the things we did when we did the FreeBSD work
is make it so that if you see a duplicate packet you send the
SACK to an alternate.. i.e. not return it to where the Data Chunk came
from but pick an alternate.. We also do the same thing once
the destination (that was the src of the Data chunk) is marked
unreachable (by hitting its threshold). 

Now I would be opposed to making this a requirement... It can
be a smart thing to do if you are implementing but I don't see
the need for it to be a requriment...

In fact there may well be other reasons which we have not thought
about where you WOULD NOT want to do what I did in our implementation.

R



"Armando L. Caro Jr." wrote:
> 
> Hi Mohammad,
> 
> I haven't followed this thread carefully. So, if I have missed something
> let me know.
> 
> I think there is a flaw in your argument. If A is down, the peer may
> notice that HBs are not being acked, but the peer may not realize yet that
> A is down. So until Path.Max.Retrans is reached, the peer still
> considers A to be up. Hence, the peer may very well send the SACK to A.
> 
> Now, if Path.Max.Retrans has been reached, then the peer DOES knows that A
> is down. However, as Jacob pointed out, you are attempting to add an
> additional requirement to the protocol by requiring that the peer send the
> SACK to B.
> 
> ~armando
> 
> 0--                                                --0
> | Armando L. Caro Jr.     |       acaro@cis.udel.edu |
> | University of Delaware  |  www.cis.udel.edu/~acaro |
> 0--                                                --0
> 
> On Wed, 24 Apr 2002, Khan, Mohammad wrote:
> 
> > Hi Jacob,
> > As u said "you cannot expect your peer to know that A is down and therefore
> > expect
> > him to send the SACK to B". Through heartbeats peer is knowing that A is
> > down and what extra it has to do is to check the status of the address on
> > which it is sending SACK if it is down it has to send the SACK to another
> > address of the same association which is alive. As far as piggybacking the
> > data with sack is concerned offcourse sending SCTP shall send it to the
> > locally know alive address of the peer.
> > Regards
> > kashif
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Wed Apr 24 17:32:22 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA09920
	for <tsvwg-archive@odin.ietf.org>; Wed, 24 Apr 2002 17:32:22 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA02794
	for tsvwg-archive@odin.ietf.org; Wed, 24 Apr 2002 17:32:26 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA01755;
	Wed, 24 Apr 2002 17:16:47 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA01723
	for <tsvwg@ns.ietf.org>; Wed, 24 Apr 2002 17:16:44 -0400 (EDT)
Received: from mail.eecis.udel.edu (louie.udel.edu [128.4.40.12])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id RAA09679
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 17:16:39 -0400 (EDT)
Received: from ren.eecis.udel.edu by mail.eecis.udel.edu id aa12686;
          24 Apr 2002 17:16 EDT
Date: Wed, 24 Apr 2002 17:16:04 -0400 (EDT)
From: "Armando L. Caro Jr." <acaro@mail.eecis.udel.edu>
To: Randall Stewart <randall@stewart.chicago.il.us>
cc: "Armando L. Caro Jr." <acaro@mail.eecis.udel.edu>,
        "Khan, Mohammad" <mo_khan@trillium.com>,
        "'Jacob Heitz'" <jheitz@lucent.com>,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] Source IP address Control
In-Reply-To: <3CC718F0.C5D186FA@stewart.chicago.il.us>
Message-ID: <Pine.GSO.4.33.0204241712350.838-100000@ren.eecis.udel.edu>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

On Wed, 24 Apr 2002, Randall Stewart wrote:

> I don't see this as a needed requirement... it is just an implementation
> decision on how they wish to handle SACK's.

I just thought from what I read that they wanted the peer to act a
particular way based on what they were doing. Expecting certain behavior
from your peer is basically making it a requirement. But like I said
before, I didn't read the thread carefully... so I may have missed
something.

> Now I would be opposed to making this a requirement... It can
> be a smart thing to do if you are implementing but I don't see
> the need for it to be a requriment...

I agree.

> In fact there may well be other reasons which we have not thought
> about where you WOULD NOT want to do what I did in our implementation.

maybe.

~armando

0--                                                --0
| Armando L. Caro Jr.     |       acaro@cis.udel.edu |
| University of Delaware  |  www.cis.udel.edu/~acaro |
0--                                                --0


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Wed Apr 24 18:47:55 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA10821
	for <tsvwg-archive@odin.ietf.org>; Wed, 24 Apr 2002 18:47:55 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id SAA06421
	for tsvwg-archive@odin.ietf.org; Wed, 24 Apr 2002 18:47:56 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA04993;
	Wed, 24 Apr 2002 18:22:13 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA04964
	for <tsvwg@ns.ietf.org>; Wed, 24 Apr 2002 18:22:10 -0400 (EDT)
Received: from ganymede.or.intel.com (jffdns01.or.intel.com [134.134.248.3])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA10459
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 18:22:05 -0400 (EDT)
Received: from orsmsxvs040.jf.intel.com (orsmsxvs040.jf.intel.com [192.168.65.206])
	by ganymede.or.intel.com (8.11.6/8.11.6/d: solo.mc,v 1.34 2002/04/15 17:47:41 root Exp $) with SMTP id g3OMLch29180
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 22:21:39 GMT
Received: from orsmsx26.jf.intel.com ([192.168.65.26])
 by orsmsxvs040.jf.intel.com (NAVGW 2.5.1.16) with SMTP id M2002042415285223247
 ; Wed, 24 Apr 2002 15:28:52 -0700
Received: by orsmsx26.jf.intel.com with Internet Mail Service (5.5.2653.19)
	id <JRN5Y1DR>; Wed, 24 Apr 2002 15:21:37 -0700
Message-ID: <F1CE15E08172D4119247009027AE9D501911739B@FMSMSX37>
From: "Balani, Sandeep" <s_balani@trillium.com>
To: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        Ivan.Arias-Rodriguez@nokia.com
Cc: "Balani, Sandeep" <s_balani@trillium.com>, qxie1@email.mot.com,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Wed, 24 Apr 2002 15:21:31 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


Hi Randy, 

" No if E2 exists and has bound IPC/IPB your SHOULD NOT ALLOW IPC to be
added to A1E1... E1 does NOT have IPC bound.

The basic concept is that a ENDPOINT must BIND the address FIRST
BEFORE you can use the dynamic IP address add. If the IP address
exists in another ENDPOINT then it is the same as if you try
to bind an address that someone else has... i.e. you get an error.

In the case of binding all ... there is no problem. you always
get all IP addresses on the machine... But once you start subset
binding the ENDPOINT must bind the address BEFORE you can add it
to ANY of the ENDPOINTS associations..."


[ Sandeep ] I like this.  

Let me understand it, When u ADDIP address on an association, U add the new
IP address on then ENDPOINT first & then Add new IP address on the
Association which was specified by the user.  Correct ?

E1 = [ IPa, IPb, Port-X]
A1 = [ E1, IPz, Port-P1 ] 
A2 = [ E1, IPy, Port-P1 ]

For Adding IPc address on association A1, First Modify endpoint as follows :


ADDIP  [ IPc]
E1 = [ IPa, IPb, IPc, Port-X] 
A1 = [ E1, IPz, Port-P1 ] ( ASCONF for Adding IPc ) 
A2 = [ E1, IPy, Port-P1 ] ( This association will only be using IPa, IPb )

But when u delete an IP address from an Association U will ONLY Delete IP
Address, from Association NOT from ENDPOINT.  Correct ?? 

Now When I do a DELETEIP of IPc from  Association A1 :  

DeleteIP [ IPc] 

E1 = [ IPa, IPb,IPc, Port-X]  ( No Change in Endpoint )
A1 = [ E1, IPz, Port-P1 ]  ( ASCONF for Deleting IPc )
A2 = [ E1, IPy, Port-P1 ]  ( No Change this assoc is using IPa, IPb ) 


Is the above example correct ??

thanks
sandeep 
 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Apr 25 02:19:57 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA27522
	for <tsvwg-archive@odin.ietf.org>; Thu, 25 Apr 2002 02:19:57 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id CAA09969
	for tsvwg-archive@odin.ietf.org; Thu, 25 Apr 2002 02:19:58 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA08764;
	Thu, 25 Apr 2002 01:57:13 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA08737
	for <tsvwg@optimus.ietf.org>; Thu, 25 Apr 2002 01:57:10 -0400 (EDT)
Received: from caduceus.fm.intel.com (fmr02.intel.com [192.55.52.25])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA18900
	for <tsvwg@ietf.org>; Thu, 25 Apr 2002 01:57:08 -0400 (EDT)
Received: from talaria.fm.intel.com (talaria.fm.intel.com [10.1.192.39])
	by caduceus.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.41 2002/04/22 23:29:17 root Exp $) with ESMTP id g3P5uLW14297
	for <tsvwg@ietf.org>; Thu, 25 Apr 2002 05:56:21 GMT
Received: from fmsmsxvs041.fm.intel.com (fmsmsxv041-1.fm.intel.com [132.233.48.109])
	by talaria.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.16 2002/04/15 17:47:00 root Exp $) with SMTP id g3P5xHx13870
	for <tsvwg@ietf.org>; Thu, 25 Apr 2002 05:59:17 GMT
Received: from FMSMSX017.fm.intel.com ([132.233.42.196])
 by fmsmsxvs041.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002042422572605096
 ; Wed, 24 Apr 2002 22:57:26 -0700
Received: by fmsmsx017.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <28FP0BMF>; Wed, 24 Apr 2002 22:57:05 -0700
Message-ID: <A1171C2850D8D4119D7B0002A508C764016983C2@bgsmsx91.iind.intel.com>
From: "Khan, Mohammad" <mo_khan@trillium.com>
To: "'Arun Prasad'" <arun@netlab.hcltech.com>,
        "Balani, Sandeep" <s_balani@trillium.com>
Cc: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        "'Qiaobing Xie'"<qxie1@email.mot.com>,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Wed, 24 Apr 2002 22:54:49 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi Arun,
Now to control ur source IP address considering the case for generic
solution across all the OS where few OS do not guarantee whether the header
source address shall remain same or not. In this scenario SCTP has to
maintain the status of all local interface addresses. Even if SCTP does so
should SCTP take decision on Association Status based on this local
interface address status.
Regards
kashif

-----Original Message-----
From: Arun Prasad [mailto:arun@netlab.hcltech.com]
Sent: Wednesday, April 24, 2002 10:28 AM
To: Balani, Sandeep
Cc: 'Randall Stewart'; 'Qiaobing Xie'; Khan, Mohammad; tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict


hi sandeep,

"Balani, Sandeep" wrote:

> Thanks Randy & Qiaobing,
>   Although i would like to add something here.
>
> So the inference of this discussion is : The source IP Addresses in an
> association can be subset of Endpoint addresses.   Is this statment
correct
> ?
>
> In the above case an Implmentation "MUST" make sure that whatever source
IP
> addresses are communicated during association intialisation should be used
> while sending outoing packets. ( I.e. control U'r Src IP Addresses )
>

    Yes thats true, but its not always that easy to implement ie., if we
implement
a generalised SCTP stack, covering many OS... then its really tough to keep
track
of this feature (full control of Src IP Addresses) in the OS....
for example:
    Windows 2000 supports IP_HDR_INCL option ( ie., adding our own Source
Ip Addresses) in its raw socket implementation, but it doesnt guarentee that
the
Src address filled by the user will be going in the IP Packet sent from that
machine...
In this case the subset concept will fail.....
But for Linux, the OS guarentees the source Ip Address given by the user
will be
filled in the IP packet sent.....
So Its more of the OS dependent... (and the OS can change this
implementation
any time....) ... The thing is, we should have a generalised API for Raw
sockets, as
we have for other protocols (like TCP, UDP etc...)

>
> Also to extend the same logic, In the example I have shown below If i
> terminate Association A1, I should also delete the entry of IP-C address
in
> Endpoint E1 Correct ?????
>



>
> Lastly this may never be TRUE but I would like to ask this question
anyways.
> In the example below I did'nt communicated IP-C on Association A2, But
Lets
> take a wild case that somehow I receive a SCTP packet on [ IP-C, Port-X,
> D2, Port-P2] then SCTP will accept this Packet Is'nt it ?  As u will find
> endpoint valid & Association  also .
>

    The Receiving endpoint will search the Association for which the packet
has
come with the following parameters { SrcAddr, SrcPort, RemAddr, RemPort}
In the above case since the SrcAddr is a different one (that is not in the
Association's
pool) the received packet will not hit on the Association A2...
So no probs....


Thanks
-arun

>
> Thanks
> Sandeep
>
> -----Original Message-----
> From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: Tuesday, April 23, 2002 5:46 PM
> To: Balani, Sandeep
> Cc: 'Qiaobing Xie'; Khan, Mohammad; tsvwg@ietf.org
> Subject: Re: [Tsvwg] End point and association Address conflict
>
> The same scenario could exist outside of addip..
>
> If I bind-all addresses.
>
> setup an assocation (with IP-A/IP-B .. the ones I have)
>
> now we hot-plug in a card.
>
> I now setup a second association with all the
> addresses I have (now IP-A/IP-B and IP-C).
>
> You end up in the same place.. .and you must,
> as Qiaobing said, assure that you only use IP addresses
> that the peer knows about for sources..
>
> Our ADDIP/DEL-IP allows this to happen as well.. we make
> it so the user has to disable auto-asconf though to reach this state.
>
> R
>
> "Balani, Sandeep" wrote:
> >
> > Hi Qiaobing,
> >     I think the reason for asking such question was we have ADDIP &
> DELETEIP
> > procedure on an Association.
> >
> > Now Assume we have an Endpoint
> >
> > E1 =  [ IP-A, IP-B, and Port-X ] .
> >
> > Now lets say i have 2 associations on endpoint E1.
> >
> > A1 = [ Endpoint-E1 , DstAddr= D1, DstPort-P1 ]
> > A2 = [ Endpoint-E2, DstAddr=D2,  DstPost-P2 ]
> >
> > Now when i do an  ADDIP [ for Address IP-C ] on association A1, actually
i
> > am adding a source  IP address in Endpoint E1 i.e. [ IP-A, IP-B, IP-C,
> > Port-X ], so that I will find an Endpoint for an incoming packet.
> >
> > But IP-C is only used by Association A1, Not by A2.  That's the reason
we
> > have an Association using Subset of IP addresses from an Endpoint.
> >
> > My Question is : Is it the intended behaviour in ADDIP/DELETEIP case
????
> >
> > Many thanks
> > sandeep
> > -----Original Message-----
> > From: Qiaobing Xie [mailto:qxie1@email.mot.com]
> > Sent: Tuesday, April 23, 2002 7:50 AM
> > To: Khan, Mohammad
> > Cc: tsvwg@ietf.org; Pillai, Hari
> > Subject: Re: [Tsvwg] End point and association Address conflict
> >
> > Hi, Kashif,
> >
> > Yes, you are allowed to do so, as long as you can guarantee that this
> > endpoint will never use Z as the source IP address in any packet sent
> > over this association. This is not hard to prevent if the endpoint has
> > full control over the source IP address selection routine.
> >
> > -Qiaobing
> >
> > "Khan, Mohammad" wrote:
> > >
> > > Hi All,
> > > I have following querry regarding endpoint and association source
> > addresses.
> > >
> > > Suppose an Endpoint has three addresses X, Y and Z.
> > >         An association which uses this endpoint and specifies only X
and
> Y
> > > addresses as source IP addresses. Is it allowed?
> > >
> > > Regards
> > > kashif
> > >
> > > _______________________________________________
> > > tsvwg mailing list
> > > tsvwg@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/tsvwg
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
>
> --
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Apr 25 07:45:12 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA01697
	for <tsvwg-archive@odin.ietf.org>; Thu, 25 Apr 2002 07:45:11 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA28923
	for tsvwg-archive@odin.ietf.org; Thu, 25 Apr 2002 07:45:13 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA26775;
	Thu, 25 Apr 2002 07:23:00 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA26748
	for <tsvwg@optimus.ietf.org>; Thu, 25 Apr 2002 07:22:57 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA01313
	for <tsvwg@ietf.org>; Thu, 25 Apr 2002 07:22:53 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3PBMIH48989;
	Thu, 25 Apr 2002 06:22:19 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CC7E6E9.862CF660@stewart.chicago.il.us>
Date: Thu, 25 Apr 2002 06:22:18 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: "Balani, Sandeep" <s_balani@trillium.com>
CC: Ivan.Arias-Rodriguez@nokia.com, qxie1@email.mot.com,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F1CE15E08172D4119247009027AE9D501911739B@FMSMSX37>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

"Balani, Sandeep" wrote:
> 
> Hi Randy,
> 
> " No if E2 exists and has bound IPC/IPB your SHOULD NOT ALLOW IPC to be
> added to A1E1... E1 does NOT have IPC bound.
> 
> The basic concept is that a ENDPOINT must BIND the address FIRST
> BEFORE you can use the dynamic IP address add. If the IP address
> exists in another ENDPOINT then it is the same as if you try
> to bind an address that someone else has... i.e. you get an error.
> 
> In the case of binding all ... there is no problem. you always
> get all IP addresses on the machine... But once you start subset
> binding the ENDPOINT must bind the address BEFORE you can add it
> to ANY of the ENDPOINTS associations..."
> 
> [ Sandeep ] I like this.

ok :>

> 
> Let me understand it, When u ADDIP address on an association, U add the new
> IP address on then ENDPOINT first & then Add new IP address on the
> Association which was specified by the user.  Correct ?
> 
> E1 = [ IPa, IPb, Port-X]
> A1 = [ E1, IPz, Port-P1 ]
> A2 = [ E1, IPy, Port-P1 ]

Yes.. the endpoint MUST gain the IP address BEFORE any of its
associations.
I guess one could do it at the same time too...

> 
> For Adding IPc address on association A1, First Modify endpoint as follows :
> 
> ADDIP  [ IPc]
> E1 = [ IPa, IPb, IPc, Port-X]
> A1 = [ E1, IPz, Port-P1 ] ( ASCONF for Adding IPc )
> A2 = [ E1, IPy, Port-P1 ] ( This association will only be using IPa, IPb )
> 
> But when u delete an IP address from an Association U will ONLY Delete IP
> Address, from Association NOT from ENDPOINT.  Correct ??
> 
I suppose you COULD allow the endpoint to remove an address
BUT you would first probably want all of its associations to
remove the IP address... or all at the same time :>

> Now When I do a DELETEIP of IPc from  Association A1 :
> 
> DeleteIP [ IPc]
> 
> E1 = [ IPa, IPb,IPc, Port-X]  ( No Change in Endpoint )
> A1 = [ E1, IPz, Port-P1 ]  ( ASCONF for Deleting IPc )
> A2 = [ E1, IPy, Port-P1 ]  ( No Change this assoc is using IPa, IPb )
> 
> Is the above example correct ??

Yes, that would be ok

R

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Thu Apr 25 18:28:05 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA08083
	for <tsvwg-archive@odin.ietf.org>; Thu, 25 Apr 2002 18:28:04 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id SAA16049
	for tsvwg-archive@odin.ietf.org; Thu, 25 Apr 2002 18:28:09 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA14957;
	Thu, 25 Apr 2002 18:07:13 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA14926
	for <tsvwg@ns.ietf.org>; Thu, 25 Apr 2002 18:07:09 -0400 (EDT)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA07642
	for <tsvwg@ietf.org>; Thu, 25 Apr 2002 18:07:04 -0400 (EDT)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir02nok.ntc.nokia.com (esvir02nokt.ntc.nokia.com [172.21.143.34])
	by mgw-x2.nokia.com (Switch-2.2.1/Switch-2.2.0) with ESMTP id g3PM7PF18844
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 01:07:25 +0300 (EET DST)
Received: from esebh003.NOE.Nokia.com (unverified) by esvir02nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5a7c39e947ac158f22078@esvir02nok.ntc.nokia.com>;
 Fri, 26 Apr 2002 01:07:06 +0300
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh003.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Fri, 26 Apr 2002 01:07:06 +0300
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Fri, 26 Apr 2002 01:07:06 +0300
Message-ID: <58AEDE27D48F884285007ABF827FF26346A39B@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] End point and association Address conflict
Thread-Index: AcHspRw/yyDKRb7AQBumkweemBPDWAAABinQ
To: <randall@stewart.chicago.il.us>
Cc: <s_balani@trillium.com>, <qxie1@email.mot.com>, <mo_khan@trillium.com>,
        <tsvwg@ietf.org>
X-OriginalArrivalTime: 25 Apr 2002 22:07:06.0660 (UTC) FILETIME=[89AC1640:01C1ECA5]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id SAA14927
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Randy, see my comment inline...

> Ivan.Arias-Rodriguez@nokia.com wrote:
> > 
> >         Hi all!
> > 
> >         Well, now I realized that I had a somehow different 
> view about what was a legal addition... thanks for your comments.
> > 
> >         Now I have another doubt... Imagine that we have an 
> endpoint E1 with addresses IPa, IPb and port 1. We have an 
> association A1 that only uses IPa, whose peer is IPz and port 
> 65535. Thus the association A1 can be represented as [(IPa; 
> Port1), (IPz; Port 65535)]. What happens now if we receive an 
> INIT from IPz/65535 directed to IPb/1?
> 
> A question...
> 
> Why did you not include IPb in the INIT or INIT-ACK that setup A1. You
> are
> supposed to tell the peer of all your addresses that are bound to your
> endpoint. I think this is the big flaw in your question..

	I did, but later on I deleted it using the AddIP extension... :-) So now I don't use it anymore...

	BR Iván Arias Rodríguez

> You really MUST tell the peer of all your endpoint's 
> addresses.  I don't
> think this is a valid scenario if you follow the proper endpoint
> procedures
> and tell your peer ALL your bound addresses..
> 
> R
> 
> 
> > 
> >         As IPb is not being used in A1, we should answer 
> with an INIT ACK as in a normal establishment of an 
> association. However, when you receive an INIT chunk, you 
> include all the addresses of the endpoint in the INIT ACK 
> chunk... But we have to be careful, because in this case, 
> when receiving the COOKIE ECHO, we have two options:
> > 
> >         a) It is directed to IPb. From our point of view, 
> it comes from an unknown peer, and then we should make a new 
> association. The problem is that we will create a new 
> association A2 that will use IPa as well, and this would be 
> in conflict with association A1, since the peer is the same one.
> > 
> >         b) It goes to IPa. Then, it will be treated as a 
> duplicate COOKIE ECHO. The local tag and the peer's tag 
> should not match the existing TCB, and the tie tags will be 
> set to 0. Thus, we are not in any of the cases that Table 2 
> considers, and we will silently discard the datagram. The 
> conclusion of this is that if the peer established an 
> association and we delete one of our source addresses, if the 
> peer actually restarts, we will be discarding the datagram 
> all the time and it will never be able to restart. Even if we 
> send DATA or HEARTBEAT, the peer will not send us back an 
> ABORT because they won't be OOTB datagrams...
> > 
> >         As when you create a new association you always 
> include all the address of the endpoint, there is no way 
> there could be another association A2 [(IPb; Port 1), (IPz; 
> Port 65535)] (unless you recognize this case when receiving 
> the INIT and you only include IPb in the INIT ACK, which I 
> don't think is a good idea at all), so I think that this 
> situation should be handled as a restart in any case... i.e., 
> if a COOKIE ECHO comes from an unknown peer, check inside the 
> State Cookie to see if we are in case a) below... However, I 
> don't know what should be done in the b) case... :-/
> > 
> >         What do the people of the list think about this?
> > 
> >         Thanks!
> > 
> >         BR Iván Arias Rodríguez
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Thu Apr 25 18:28:34 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA08103
	for <tsvwg-archive@odin.ietf.org>; Thu, 25 Apr 2002 18:28:30 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id SAA16064
	for tsvwg-archive@odin.ietf.org; Thu, 25 Apr 2002 18:28:32 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA14793;
	Thu, 25 Apr 2002 18:04:18 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA14764
	for <tsvwg@ns.ietf.org>; Thu, 25 Apr 2002 18:04:15 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA07609
	for <tsvwg@ietf.org>; Thu, 25 Apr 2002 18:04:09 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3PM3tH49902;
	Thu, 25 Apr 2002 17:03:55 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CC87D4A.18177211@stewart.chicago.il.us>
Date: Thu, 25 Apr 2002 17:03:55 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: s_balani@trillium.com, qxie1@email.mot.com, mo_khan@trillium.com,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFAE@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Ivan.Arias-Rodriguez@nokia.com wrote:
> 
>         Hi all!
> 
>         Well, now I realized that I had a somehow different view about what was a legal addition... thanks for your comments.
> 
>         Now I have another doubt... Imagine that we have an endpoint E1 with addresses IPa, IPb and port 1. We have an association A1 that only uses IPa, whose peer is IPz and port 65535. Thus the association A1 can be represented as [(IPa; Port1), (IPz; Port 65535)]. What happens now if we receive an INIT from IPz/65535 directed to IPb/1?

A question...

Why did you not include IPb in the INIT or INIT-ACK that setup A1. You
are
supposed to tell the peer of all your addresses that are bound to your
endpoint. I think this is the big flaw in your question..

You really MUST tell the peer of all your endpoint's addresses.  I don't
think this is a valid scenario if you follow the proper endpoint
procedures
and tell your peer ALL your bound addresses..

R


> 
>         As IPb is not being used in A1, we should answer with an INIT ACK as in a normal establishment of an association. However, when you receive an INIT chunk, you include all the addresses of the endpoint in the INIT ACK chunk... But we have to be careful, because in this case, when receiving the COOKIE ECHO, we have two options:
> 
>         a) It is directed to IPb. From our point of view, it comes from an unknown peer, and then we should make a new association. The problem is that we will create a new association A2 that will use IPa as well, and this would be in conflict with association A1, since the peer is the same one.
> 
>         b) It goes to IPa. Then, it will be treated as a duplicate COOKIE ECHO. The local tag and the peer's tag should not match the existing TCB, and the tie tags will be set to 0. Thus, we are not in any of the cases that Table 2 considers, and we will silently discard the datagram. The conclusion of this is that if the peer established an association and we delete one of our source addresses, if the peer actually restarts, we will be discarding the datagram all the time and it will never be able to restart. Even if we send DATA or HEARTBEAT, the peer will not send us back an ABORT because they won't be OOTB datagrams...
> 
>         As when you create a new association you always include all the address of the endpoint, there is no way there could be another association A2 [(IPb; Port 1), (IPz; Port 65535)] (unless you recognize this case when receiving the INIT and you only include IPb in the INIT ACK, which I don't think is a good idea at all), so I think that this situation should be handled as a restart in any case... i.e., if a COOKIE ECHO comes from an unknown peer, check inside the State Cookie to see if we are in case a) below... However, I don't know what should be done in the b) case... :-/
> 
>         What do the people of the list think about this?
> 
>         Thanks!
> 
>         BR Iván Arias Rodríguez

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Thu Apr 25 18:33:57 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA08259
	for <tsvwg-archive@odin.ietf.org>; Thu, 25 Apr 2002 18:33:57 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id SAA16785
	for tsvwg-archive@odin.ietf.org; Thu, 25 Apr 2002 18:34:01 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA14042;
	Thu, 25 Apr 2002 17:58:31 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA14014
	for <tsvwg@ns.ietf.org>; Thu, 25 Apr 2002 17:58:28 -0400 (EDT)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA07502
	for <tsvwg@ietf.org>; Thu, 25 Apr 2002 17:58:23 -0400 (EDT)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir04nok.ntc.nokia.com (esvir04nokt.ntc.nokia.com [172.21.143.36])
	by mgw-x2.nokia.com (Switch-2.2.1/Switch-2.2.0) with ESMTP id g3PLwiF17172
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 00:58:44 +0300 (EET DST)
Received: from esebh003.NOE.Nokia.com (unverified) by esvir04nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5a7c31f44eac158f24078@esvir04nok.ntc.nokia.com>;
 Fri, 26 Apr 2002 00:58:25 +0300
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh003.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Fri, 26 Apr 2002 00:58:25 +0300
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Fri, 26 Apr 2002 00:58:24 +0300
Message-ID: <58AEDE27D48F884285007ABF827FF26308EFAE@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] End point and association Address conflict
Thread-Index: AcHsS3+O9fBE4p9gQc6xJFsHeUJUEAAVJbqA
To: <randall@stewart.chicago.il.us>, <s_balani@trillium.com>
Cc: <qxie1@email.mot.com>, <mo_khan@trillium.com>, <tsvwg@ietf.org>
X-OriginalArrivalTime: 25 Apr 2002 21:58:25.0165 (UTC) FILETIME=[52D643D0:01C1ECA4]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id RAA14015
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi all!

	Well, now I realized that I had a somehow different view about what was a legal addition... thanks for your comments.

	Now I have another doubt... Imagine that we have an endpoint E1 with addresses IPa, IPb and port 1. We have an association A1 that only uses IPa, whose peer is IPz and port 65535. Thus the association A1 can be represented as [(IPa; Port1), (IPz; Port 65535)]. What happens now if we receive an INIT from IPz/65535 directed to IPb/1?

	As IPb is not being used in A1, we should answer with an INIT ACK as in a normal establishment of an association. However, when you receive an INIT chunk, you include all the addresses of the endpoint in the INIT ACK chunk... But we have to be careful, because in this case, when receiving the COOKIE ECHO, we have two options:

	a) It is directed to IPb. From our point of view, it comes from an unknown peer, and then we should make a new association. The problem is that we will create a new association A2 that will use IPa as well, and this would be in conflict with association A1, since the peer is the same one.

	b) It goes to IPa. Then, it will be treated as a duplicate COOKIE ECHO. The local tag and the peer's tag should not match the existing TCB, and the tie tags will be set to 0. Thus, we are not in any of the cases that Table 2 considers, and we will silently discard the datagram. The conclusion of this is that if the peer established an association and we delete one of our source addresses, if the peer actually restarts, we will be discarding the datagram all the time and it will never be able to restart. Even if we send DATA or HEARTBEAT, the peer will not send us back an ABORT because they won't be OOTB datagrams...

	As when you create a new association you always include all the address of the endpoint, there is no way there could be another association A2 [(IPb; Port 1), (IPz; Port 65535)] (unless you recognize this case when receiving the INIT and you only include IPb in the INIT ACK, which I don't think is a good idea at all), so I think that this situation should be handled as a restart in any case... i.e., if a COOKIE ECHO comes from an unknown peer, check inside the State Cookie to see if we are in case a) below... However, I don't know what should be done in the b) case... :-/

	What do the people of the list think about this?

	Thanks!

	BR Iván Arias Rodríguez

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Thu Apr 25 18:44:49 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA08378
	for <tsvwg-archive@odin.ietf.org>; Thu, 25 Apr 2002 18:44:49 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id SAA16941
	for tsvwg-archive@odin.ietf.org; Thu, 25 Apr 2002 18:44:50 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA16120;
	Thu, 25 Apr 2002 18:29:16 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA16091
	for <tsvwg@ns.ietf.org>; Thu, 25 Apr 2002 18:29:13 -0400 (EDT)
Received: from ganymede.or.intel.com (jffdns01.or.intel.com [134.134.248.3])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA08118
	for <tsvwg@ietf.org>; Thu, 25 Apr 2002 18:29:08 -0400 (EDT)
Received: from orsmsxvs040.jf.intel.com (orsmsxvs040.jf.intel.com [192.168.65.206])
	by ganymede.or.intel.com (8.11.6/8.11.6/d: solo.mc,v 1.34 2002/04/15 17:47:41 root Exp $) with SMTP id g3PMSf623798
	for <tsvwg@ietf.org>; Thu, 25 Apr 2002 22:28:41 GMT
Received: from orsmsx26.jf.intel.com ([192.168.65.26])
 by orsmsxvs040.jf.intel.com (NAVGW 2.5.1.16) with SMTP id M2002042515355710208
 ; Thu, 25 Apr 2002 15:35:57 -0700
Received: by orsmsx26.jf.intel.com with Internet Mail Service (5.5.2653.19)
	id <J4QW99QB>; Thu, 25 Apr 2002 15:28:41 -0700
Message-ID: <F1CE15E08172D4119247009027AE9D50191173A0@FMSMSX37>
From: "Balani, Sandeep" <s_balani@trillium.com>
To: "'Ivan.Arias-Rodriguez@nokia.com'" <Ivan.Arias-Rodriguez@nokia.com>,
        randall@stewart.chicago.il.us
Cc: "Balani, Sandeep" <s_balani@trillium.com>, qxie1@email.mot.com,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Thu, 25 Apr 2002 15:28:38 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id SAA16092
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

hi all, 
 
My comments inline, 
  
-----Original Message-----
From: Ivan.Arias-Rodriguez@nokia.com
[mailto:Ivan.Arias-Rodriguez@nokia.com]
Sent: Thursday, April 25, 2002 3:07 PM
To: randall@stewart.chicago.il.us
Cc: s_balani@trillium.com; qxie1@email.mot.com; mo_khan@trillium.com;
tsvwg@ietf.org
Subject: RE: [Tsvwg] End point and association Address conflict


	Randy, see my comment inline...

> Ivan.Arias-Rodriguez@nokia.com wrote:
> > 
> >         Hi all!
> > 
> >         Well, now I realized that I had a somehow different 
> view about what was a legal addition... thanks for your comments.
> > 
> >         Now I have another doubt... Imagine that we have an 
> endpoint E1 with addresses IPa, IPb and port 1. We have an 
> association A1 that only uses IPa, whose peer is IPz and port 
> 65535. Thus the association A1 can be represented as [(IPa; 
> Port1), (IPz; Port 65535)]. What happens now if we receive an 
> INIT from IPz/65535 directed to IPb/1?
> 
> A question...
> 
> Why did you not include IPb in the INIT or INIT-ACK that setup A1. You
> are
> supposed to tell the peer of all your addresses that are bound to your
> endpoint. I think this is the big flaw in your question..

	I did, but later on I deleted it using the AddIP extension... :-) So
now I don't use it anymore...

[ sandeep ] I think this is very interesting proble :-)
 What i think we can do is When u are doing DELETEIP for IPb, If no
association is using IPb as source address we should remover IPb from
Endpoint E1. This way we will not have any such problem. 

What do u think ????

	BR Iván Arias Rodríguez

> You really MUST tell the peer of all your endpoint's 
> addresses.  I don't
> think this is a valid scenario if you follow the proper endpoint
> procedures
> and tell your peer ALL your bound addresses..
> 
> R
> 
> 
> > 
> >         As IPb is not being used in A1, we should answer 
> with an INIT ACK as in a normal establishment of an 
> association. However, when you receive an INIT chunk, you 
> include all the addresses of the endpoint in the INIT ACK 
> chunk... But we have to be careful, because in this case, 
> when receiving the COOKIE ECHO, we have two options:
> > 
> >         a) It is directed to IPb. From our point of view, 
> it comes from an unknown peer, and then we should make a new 
> association. The problem is that we will create a new 
> association A2 that will use IPa as well, and this would be 
> in conflict with association A1, since the peer is the same one.
> > 
> >         b) It goes to IPa. Then, it will be treated as a 
> duplicate COOKIE ECHO. The local tag and the peer's tag 
> should not match the existing TCB, and the tie tags will be 
> set to 0. Thus, we are not in any of the cases that Table 2 
> considers, and we will silently discard the datagram. The 
> conclusion of this is that if the peer established an 
> association and we delete one of our source addresses, if the 
> peer actually restarts, we will be discarding the datagram 
> all the time and it will never be able to restart. Even if we 
> send DATA or HEARTBEAT, the peer will not send us back an 
> ABORT because they won't be OOTB datagrams...
> > 
> >         As when you create a new association you always 
> include all the address of the endpoint, there is no way 
> there could be another association A2 [(IPb; Port 1), (IPz; 
> Port 65535)] (unless you recognize this case when receiving 
> the INIT and you only include IPb in the INIT ACK, which I 
> don't think is a good idea at all), so I think that this 
> situation should be handled as a restart in any case... i.e., 
> if a COOKIE ECHO comes from an unknown peer, check inside the 
> State Cookie to see if we are in case a) below... However, I 
> don't know what should be done in the b) case... :-/
> > 
> >         What do the people of the list think about this?
> > 
> >         Thanks!
> > 
> >         BR Iván Arias Rodríguez
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Apr 26 01:24:48 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA16775
	for <tsvwg-archive@odin.ietf.org>; Fri, 26 Apr 2002 01:24:48 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id BAA14610
	for tsvwg-archive@odin.ietf.org; Fri, 26 Apr 2002 01:24:50 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA05199;
	Fri, 26 Apr 2002 01:01:27 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA05158
	for <tsvwg@ns.ietf.org>; Fri, 26 Apr 2002 01:01:23 -0400 (EDT)
Received: from mail-gw.radisys.com (mail-gw.radisys.com [206.102.10.35])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA16475
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 01:01:19 -0400 (EDT)
From: Dave.Sclarsky@radisys.com
Received: by mail-gw.radisys.com (Postfix, from userid 5)
	id 59D984E88A; Thu, 25 Apr 2002 22:01:20 -0700 (PDT)
Received: from notes.radisys.com(206.103.52.220)
 via SMTP by mail-gw.radisys.com, id smtpdAAA0HaWpD; Thu Apr 25 22:01:00 2002
To: tsvwg@ietf.org
Cc: Simon.Zhuang@radisys.com, Heinz.Prantner@radisys.com,
        Gene.Litt@radisys.com, Seamus.Gilchrist@radisys.com
X-Mailer: Lotus Notes Release 5.0.7  March 21, 2001
Message-ID: <OFE3B784B7.884E9280-ON85256BA7.0016B1AC@radisys.com>
Date: Fri, 26 Apr 2002 00:47:34 -0400
X-MIMETrack: Serialize by Router on HQ_ACS_1/Radisys_Corporation/US(Release 5.0.9a |January
 7, 2002) at 04/25/2002 09:52:49 PM
MIME-Version: 1.0
Content-type: text/plain; charset=us-ascii
Subject: [Tsvwg] T3 cannot timeout???
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hello,
We've found a Here's a case we found, hopefully someone can help!

2 peers (A and B) with 2 destinations (1 and 2) each, both using the same
path as the 'primary'.
Primary ethernet goes down.
Side A detects this, and switches to destination 2.
Side B's T3 timer hasn't yet expired, but starts receiving data from A on
2.  B sends SACK on 2nd path (since that's where it got the data), and
since it also has data to send, it bundles DATA chunks with the SACK.
The data bundled with the (many?) SACKs on dest 2 will eventually be
SACKed.  These SACKs must contain GAP reports, indicating the missing TSNs
(the ones sent to dest 1 with TSNs lower than those bundled with the
SACKs).  Once 4 of these SACKs with GAPs are received, Side B will "fast
retransmit" those chunks on dest 2.  According to the RFC (7.2.4), Side B
will restart T3 for dest 1, since it will be retransmitting the lowest TSN
sent to it.  It appears that this will continue forever.  Side B will never
see dest 1 as failed, since its T3 will continually be reset, but it will
waste a lot of time and energy sending it data!

Has anyone seen this?  Any suggestions?

======================================
Dave Sclarsky
Radisys Corporation
520 Fellowship Road
Suite C-304
Mt. Laurel, NJ 08054
email:dave.sclarsky@radisys.com
voice:856.642.9229x231 fax:856.642.0252



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Apr 26 04:56:51 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA27643
	for <tsvwg-archive@odin.ietf.org>; Fri, 26 Apr 2002 04:56:51 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id EAA25197
	for tsvwg-archive@odin.ietf.org; Fri, 26 Apr 2002 04:56:54 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA23904;
	Fri, 26 Apr 2002 04:30:21 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA23849
	for <tsvwg@optimus.ietf.org>; Fri, 26 Apr 2002 04:30:16 -0400 (EDT)
Received: from mgw-x3.nokia.com (mgw-x3.nokia.com [131.228.20.26])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA27197
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 04:30:07 -0400 (EDT)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir01nok.ntc.nokia.com (esvir01nokt.ntc.nokia.com [172.21.143.33])
	by mgw-x3.nokia.com (Switch-2.2.1/Switch-2.2.0) with ESMTP id g3Q8VIJ12112
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 11:31:18 +0300 (EET DST)
Received: from esebh003.NOE.Nokia.com (unverified) by esvir01nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5a7e744a1fac158f2113d@esvir01nok.ntc.nokia.com>;
 Fri, 26 Apr 2002 11:30:07 +0300
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh003.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Fri, 26 Apr 2002 11:30:06 +0300
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Fri, 26 Apr 2002 11:30:06 +0300
Message-ID: <58AEDE27D48F884285007ABF827FF26308EFAF@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] End point and association Address conflict
Thread-Index: AcHsvRB6XKDhvrXJRkOWKQjMPWIVpAAPO90Q
To: <qxie1@email.mot.com>
Cc: <randall@stewart.chicago.il.us>, <s_balani@trillium.com>,
        <Qiaobing.Xie@motorola.com>, <mo_khan@trillium.com>, <tsvwg@ietf.org>
X-OriginalArrivalTime: 26 Apr 2002 08:30:06.0975 (UTC) FILETIME=[921380F0:01C1ECFC]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id EAA23867
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi all!

	I think I haven't been so clear when explaining my doubt... Let me make it clearer...

> Ivan.Arias-Rodriguez@nokia.com wrote:
> ....
> >         Now I have another doubt... Imagine that we have an 
> endpoint 
> > E1 with addresses IPa, IPb and port 1. We have an association A1 
> > that only uses IPa, whose peer is IPz and port 65535. Thus the 
> > association A1 can be represented as [(IPa; Port1), (IPz; 
> Port 65535)]. 
> > What happens now if we receive an INIT from IPz/65535 
> directed to IPb/1?
> > 
> >         As IPb is not being used in A1, we should answer with an 
> > INIT ACK as in a normal establishment of an association. However, 
> > when you receive an INIT chunk, you include all the 
> addresses of the endpoint
>                                       ^^^^^^^^^^^^^
> Here your "..include all the addr..." directly violated the "under no
> circumstance can an SCTP transport address ever be present in 
> more than
> one SCTP endpoint" rule (see last the paragraph, page 23 of the SCTP
> book).
> 
>         E1               E2
> A1 = [(IPa; Port1), (IPz; Port 65535)]
> 
>         E3                 E2
> A2 = [(IPa, IPb; Port1), (IPz; Port 65535)]
> 
> And you have IPa/Port1 present in both E1 and E3 (they are 
> TWO different
> endpoints by definition!). This is prohibited by the SCTP association
> definition.

	No, I don't think what I'm proposing is illegal at all... Take into account that you can make use of the AddIP extension to delete IPb... Thus, initially, the association had all the addresses of the endpoint (IPa and IPb). Later on you delete your source address IPb. And then all this happens...

	So, the problem is that the pair IPb/1-IPz/65535 is not part of any association... :-/

	BR Iván Arias Rodríguez

> -Qiaobing
> 
> > in the INIT ACK chunk... But we have to be careful, because in
> > this case, when receiving the COOKIE ECHO, we have two options:
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Apr 26 05:12:03 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA27829
	for <tsvwg-archive@odin.ietf.org>; Fri, 26 Apr 2002 05:12:03 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA26265
	for tsvwg-archive@odin.ietf.org; Fri, 26 Apr 2002 05:12:06 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA24815;
	Fri, 26 Apr 2002 04:43:36 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA24789
	for <tsvwg@optimus.ietf.org>; Fri, 26 Apr 2002 04:43:34 -0400 (EDT)
Received: from mgw-x1.nokia.com (mgw-x1.nokia.com [131.228.20.21])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA27390
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 04:43:30 -0400 (EDT)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir05nok.ntc.nokia.com (esvir05nokt.ntc.nokia.com [172.21.143.37])
	by mgw-x1.nokia.com (Switch-2.2.1/Switch-2.2.0) with ESMTP id g3Q8gYu11412
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 11:42:35 +0300 (EET DST)
Received: from esebh002.NOE.Nokia.com (unverified) by esvir05nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5a7e80933bac158f2510b7@esvir05nok.ntc.nokia.com>;
 Fri, 26 Apr 2002 11:43:32 +0300
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh002.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Fri, 26 Apr 2002 11:43:32 +0300
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Fri, 26 Apr 2002 11:43:31 +0300
Message-ID: <58AEDE27D48F884285007ABF827FF26346A39E@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] End point and association Address conflict
Thread-Index: AcHsqI9tbkWO58fuRH6vmdRTpRuxUAAUVHEw
To: <s_balani@trillium.com>, <randall@stewart.chicago.il.us>
Cc: <qxie1@email.mot.com>, <mo_khan@trillium.com>, <tsvwg@ietf.org>
X-OriginalArrivalTime: 26 Apr 2002 08:43:32.0099 (UTC) FILETIME=[71F7AD30:01C1ECFE]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id EAA24790
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi Sandeep!

	See my comment inline...

> hi all, 
>  
> My comments inline, 
>   
> 
> 	Randy, see my comment inline...
> 
> > Ivan.Arias-Rodriguez@nokia.com wrote:
> > > 
> > >         Hi all!
> > > 
> > >         Well, now I realized that I had a somehow different 
> > view about what was a legal addition... thanks for your comments.
> > > 
> > >         Now I have another doubt... Imagine that we have an 
> > endpoint E1 with addresses IPa, IPb and port 1. We have an 
> > association A1 that only uses IPa, whose peer is IPz and port 
> > 65535. Thus the association A1 can be represented as [(IPa; 
> > Port1), (IPz; Port 65535)]. What happens now if we receive an 
> > INIT from IPz/65535 directed to IPb/1?
> > 
> > A question...
> > 
> > Why did you not include IPb in the INIT or INIT-ACK that 
> setup A1. You
> > are
> > supposed to tell the peer of all your addresses that are 
> bound to your
> > endpoint. I think this is the big flaw in your question..
> 
> 	I did, but later on I deleted it using the AddIP 
> extension... :-) So
> now I don't use it anymore...
> 
> [ sandeep ] I think this is very interesting proble :-)
>  What i think we can do is When u are doing DELETEIP for IPb, If no
> association is using IPb as source address we should remover IPb from
> Endpoint E1. This way we will not have any such problem. 
> 
> What do u think ????

	No, I don't think this is a good solution... Remember that there can be many associations in that endpoint, and that they might be using IPb as A1 did once... I don't think you can erase that address from the endpoint...

	BR Iván Arias Rodríguez

> 	BR Iván Arias Rodríguez
> 
> > You really MUST tell the peer of all your endpoint's 
> > addresses.  I don't
> > think this is a valid scenario if you follow the proper endpoint
> > procedures
> > and tell your peer ALL your bound addresses..
> > 
> > R
> > 
> > 
> > > 
> > >         As IPb is not being used in A1, we should answer 
> > with an INIT ACK as in a normal establishment of an 
> > association. However, when you receive an INIT chunk, you 
> > include all the addresses of the endpoint in the INIT ACK 
> > chunk... But we have to be careful, because in this case, 
> > when receiving the COOKIE ECHO, we have two options:
> > > 
> > >         a) It is directed to IPb. From our point of view, 
> > it comes from an unknown peer, and then we should make a new 
> > association. The problem is that we will create a new 
> > association A2 that will use IPa as well, and this would be 
> > in conflict with association A1, since the peer is the same one.
> > > 
> > >         b) It goes to IPa. Then, it will be treated as a 
> > duplicate COOKIE ECHO. The local tag and the peer's tag 
> > should not match the existing TCB, and the tie tags will be 
> > set to 0. Thus, we are not in any of the cases that Table 2 
> > considers, and we will silently discard the datagram. The 
> > conclusion of this is that if the peer established an 
> > association and we delete one of our source addresses, if the 
> > peer actually restarts, we will be discarding the datagram 
> > all the time and it will never be able to restart. Even if we 
> > send DATA or HEARTBEAT, the peer will not send us back an 
> > ABORT because they won't be OOTB datagrams...
> > > 
> > >         As when you create a new association you always 
> > include all the address of the endpoint, there is no way 
> > there could be another association A2 [(IPb; Port 1), (IPz; 
> > Port 65535)] (unless you recognize this case when receiving 
> > the INIT and you only include IPb in the INIT ACK, which I 
> > don't think is a good idea at all), so I think that this 
> > situation should be handled as a restart in any case... i.e., 
> > if a COOKIE ECHO comes from an unknown peer, check inside the 
> > State Cookie to see if we are in case a) below... However, I 
> > don't know what should be done in the b) case... :-/
> > > 
> > >         What do the people of the list think about this?
> > > 
> > >         Thanks!
> > > 
> > >         BR Iván Arias Rodríguez
> > 
> > -- 
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> > 
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Apr 26 05:25:19 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA28016
	for <tsvwg-archive@odin.ietf.org>; Fri, 26 Apr 2002 05:25:14 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA26664
	for tsvwg-archive@odin.ietf.org; Fri, 26 Apr 2002 05:25:17 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA26022;
	Fri, 26 Apr 2002 05:03:52 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA25977
	for <tsvwg@optimus.ietf.org>; Fri, 26 Apr 2002 05:03:48 -0400 (EDT)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA27743
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 05:03:44 -0400 (EDT)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir02nok.ntc.nokia.com (esvir02nokt.ntc.nokia.com [172.21.143.34])
	by mgw-x2.nokia.com (Switch-2.2.1/Switch-2.2.0) with ESMTP id g3Q941F00671
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 12:04:01 +0300 (EET DST)
Received: from esebh003.NOE.Nokia.com (unverified) by esvir02nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5a7e930cefac158f22078@esvir02nok.ntc.nokia.com>;
 Fri, 26 Apr 2002 12:03:42 +0300
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh003.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Fri, 26 Apr 2002 12:03:42 +0300
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] T3 cannot timeout???
Date: Fri, 26 Apr 2002 12:03:41 +0300
Message-ID: <58AEDE27D48F884285007ABF827FF26308EFB0@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] T3 cannot timeout???
Thread-Index: AcHs33UMd0PqK2OcSkaGIsahTzkniQAHzgJg
To: <Dave.Sclarsky@radisys.com>, <tsvwg@ietf.org>
Cc: <Simon.Zhuang@radisys.com>, <Heinz.Prantner@radisys.com>,
        <Gene.Litt@radisys.com>, <Seamus.Gilchrist@radisys.com>
X-OriginalArrivalTime: 26 Apr 2002 09:03:42.0382 (UTC) FILETIME=[435A34E0:01C1ED01]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id FAA25978
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi Dave!

	I must confess that I don't really follow you, but let's try...

> Hello,
> We've found a Here's a case we found, hopefully someone can help!
> 
> 2 peers (A and B) with 2 destinations (1 and 2) each, both 
> using the same
> path as the 'primary'.

	Let me check if I know what you are saying here... Peer A uses IPa1 and IPa2, and peer B uses IPb1 and IPb2... So, you mean that A's primary address is IPb1 and uses IPa1 as the source address, and peer B uses IPa1 as primary and IPb1 as the source address for those datagrams, right?

> Primary ethernet goes down.

	Which one? IPb1?

> Side A detects this, and switches to destination 2.

	Ok, now A is using IPb2...

> Side B's T3 timer hasn't yet expired, but starts receiving 
> data from A on

	Ok... I think I understand you now that I read it again... You mean that there are two independent Ethernet networks... So IPa1 and IPb1 are in the same network, and IPa2 and IPb2 are in another network. Am I right?

> 2.  B sends SACK on 2nd path (since that's where it got the data), and
> since it also has data to send, it bundles DATA chunks with the SACK.

	I don't think this is the right thing to do... You always use the primary address until it is down... I would most probably do it in the other way: sending SACKs bundled with the DATA chunks sent to the primary address...

> The data bundled with the (many?) SACKs on dest 2 will eventually be
> SACKed.  These SACKs must contain GAP reports, indicating the 
> missing TSNs
> (the ones sent to dest 1 with TSNs lower than those bundled with the
> SACKs).  Once 4 of these SACKs with GAPs are received, Side B 
> will "fast
> retransmit" those chunks on dest 2.  According to the RFC 

	Well, that's right... However I think you shouldn't send the DATA to IPa2 just because you are also sending SACKs there...

> (7.2.4), Side B
> will restart T3 for dest 1, since it will be retransmitting 
> the lowest TSN
> sent to it.  It appears that this will continue forever.  
> Side B will never
> see dest 1 as failed, since its T3 will continually be reset, 
> but it will
> waste a lot of time and energy sending it data!

	Ok, so the problem is that you are sending DATA chunks both to IPa1 and IPa2 depending on if you send SACKs or not... I think this is the problem... T3 in IPa1 won't reset because the SACKs of the DATA chunks sent to IPa2 will always trigger a fast retransmission before it expires...

	However, if you use only the primary address when sending data, you don't have this problem... I don't know if this is the solution, but the point is that what you are doing sending DATA to the secondary address is not the right thing to do...

	BR Iván Arias Rodríguez

> Has anyone seen this?  Any suggestions?
> 
> ======================================
> Dave Sclarsky
> Radisys Corporation
> 520 Fellowship Road
> Suite C-304
> Mt. Laurel, NJ 08054
> email:dave.sclarsky@radisys.com
> voice:856.642.9229x231 fax:856.642.0252
> 
> 
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Apr 26 09:42:06 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA03989
	for <tsvwg-archive@odin.ietf.org>; Fri, 26 Apr 2002 09:42:06 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA12499
	for tsvwg-archive@odin.ietf.org; Fri, 26 Apr 2002 09:42:09 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA10957;
	Fri, 26 Apr 2002 09:23:02 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA10926
	for <tsvwg@optimus.ietf.org>; Fri, 26 Apr 2002 09:22:58 -0400 (EDT)
Received: from nef.ens.fr (nef.ens.fr [129.199.96.32])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA03270
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 09:22:53 -0400 (EDT)
Received: from baobab.ens.fr (baobab.ens.fr [129.199.99.66])
          by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id g3QDMKa52599
          ; Fri, 26 Apr 2002 15:22:20 +0200 (CEST)
Received: from localhost (hong@localhost) by baobab.ens.fr (8.9.3/jb-1.1)
	id PAA19183 ; Fri, 26 Apr 2002 15:22:19 +0200 (CEST)
Date: Fri, 26 Apr 2002 15:22:19 +0200 (CEST)
From: Dohy HONG <Dohy.Hong@ens.fr>
To: tsvwg@ietf.org, Scott Bradner <sob@harvard.edu>
Message-ID: <Pine.BSF.4.21.0204261520160.19165-100000@baobab.ens.fr>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=ISO-8859-1
Content-Transfer-Encoding: 8BIT
Content-Transfer-Encoding: 8BIT
Subject: [Tsvwg] FTCP algorithm
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8BIT


Hi all,

I would like to open discussions about the ideas of
FTCP algorithm presented in:
http://www.ietf.org/internet-drafts/draft-hong-ftcp-00.txt

This algorithm propose a simple modification of
TCP algorithm to improve the performance in a way
which is 'socially better'.
And it seems that there are 2 interesting properties
in using FTCP (F for fluid):
One is that it improve the End-to-End performance for
its user.
The second is that it encourages others to also use
this algorithm and become 'socially better'.

The basic idea is based on the remark that the
packet burstiness inside each RTT is not good:
especially with Slow Start and long RTTs.
If you have CWND=10 with RTT~1 second, it is better
to send 1 packet every 100 ms that to send a burst
of 10 packets.

What is your gain? I did some simple test modifying
the linux kernel and doing FTP and HTTP
(I also did some simple NS simulation which is described
in the draft).
My feeling was that you could gain about 0-50%
depending on RTT, congestion state etc. It would be
nice if other people could check this
independently. I would be more than happy to help
anyone who would like to do kernel modification 
(or for ns-2) for FTCP and get any feedback. 

What is the cost? Well, I think FTCP idea is more
efficient with a good timer granularity.
Timer is needed by the fact that you need to be
patient before sending packets after ACKs reception
(roughly speaking: try to send 1 packet every
CWND/RTT).
A good point is also that this idea could be
implemented at the source or/and the destination level.

I naively hope that FTCP could become useful to
people in practice.

All feedbacks, comments, criticisms (even blames) are very
welcome.
Dohy

===================================================
Dohy HONG
INRIA-ENS
Projet TREC (Théorie des Réseaux Et Communications)

ECOLE NORMALE SUPERIEURE 
Département Informatique       
45 rue d'Ulm, 75230 PARIS CEDEX 05,  FRANCE    
Tel : 33 (1) 4432 2112, Fax : 33 (1) 4432 2151
        ------------------------------------
Email : Dohy.Hong@ens.fr
http://www.di.ens.fr/~trec
===================================================


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Apr 26 09:58:17 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA04537
	for <tsvwg-archive@odin.ietf.org>; Fri, 26 Apr 2002 09:58:17 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA13140
	for tsvwg-archive@odin.ietf.org; Fri, 26 Apr 2002 09:58:19 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA11283;
	Fri, 26 Apr 2002 09:28:47 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA11200
	for <tsvwg@optimus.ietf.org>; Fri, 26 Apr 2002 09:28:37 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA03366
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 09:28:33 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3QDR5H51419;
	Fri, 26 Apr 2002 08:27:06 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CC955A9.A31DE2A6@stewart.chicago.il.us>
Date: Fri, 26 Apr 2002 08:27:05 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: s_balani@trillium.com, qxie1@email.mot.com, mo_khan@trillium.com,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <58AEDE27D48F884285007ABF827FF26346A39B@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Ivan.Arias-Rodriguez@nokia.com wrote:
> 
>         Randy, see my comment inline...
> 
> > Ivan.Arias-Rodriguez@nokia.com wrote:
> > >
> > >         Hi all!
> > >
> > >         Well, now I realized that I had a somehow different
> > view about what was a legal addition... thanks for your comments.
> > >
> > >         Now I have another doubt... Imagine that we have an
> > endpoint E1 with addresses IPa, IPb and port 1. We have an
> > association A1 that only uses IPa, whose peer is IPz and port
> > 65535. Thus the association A1 can be represented as [(IPa;
> > Port1), (IPz; Port 65535)]. What happens now if we receive an
> > INIT from IPz/65535 directed to IPb/1?
> >
> > A question...
> >
> > Why did you not include IPb in the INIT or INIT-ACK that setup A1. You
> > are
> > supposed to tell the peer of all your addresses that are bound to your
> > endpoint. I think this is the big flaw in your question..
> 
>         I did, but later on I deleted it using the AddIP extension... :-) So now I don't use it anymore...
> 

Ok, but I don't get why you would do that... AddIP is intended for
when you add and delete interfaces... Using it to "tune" your 
interface addresses to have a different binding set then
your interface is not what it is intended for. I would think
you should have subtracted from the associations AND the endpoint.

Now if you think about this, you COULD have the same thing
occur when binding all (without AddIP)... i.e. you have bound all..
setup
an association.

Then you plug in a NEW card ifconfig it up and add an address.

Now your peer of the association sends a INIT()---> to your new address
and your existing port... which now showed up in your endpoint
since you are bound to ALL addresses...

AddIP fixes this in that you can add the address when it is ifconfig'd
up... 

I think that:


A) This is a real real corner case .. why would the peer endpoint try
    to communicate with you AGAIN .. remember it MUST be the same
    socket.. .when you already have an association?

 B) If you don't allow a delete of a address from an association unless
    it is also deleted from the endpoint.. the problem goes away.. hmm
    maybe this needs to be a rule for add-ip? not sure on this.. I want
    to think about it.. maybe not.. after all endpoints vs associations
    in some ways are a implementation issue...hmm..

 C) This can occur in current SCTP when you have the ability to
dynamically
    add a card...

Hmm thinking about it somemore.. if you are bound-all to the endpoint
you
won't have this problem.. since even if you deleted the address you most
likely will FIND the association and it will then look like a restart.

It will really confuse the other endpoint though.. since his old
association
will not work quite right.. it also will find the old association when
it gets your address list... hmm...

I will have to run this one around in my head for a bit..

R

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Apr 26 09:59:40 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA04609
	for <tsvwg-archive@odin.ietf.org>; Fri, 26 Apr 2002 09:59:40 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA13185
	for tsvwg-archive@odin.ietf.org; Fri, 26 Apr 2002 09:59:43 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA11729;
	Fri, 26 Apr 2002 09:31:56 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA11700
	for <tsvwg@optimus.ietf.org>; Fri, 26 Apr 2002 09:31:52 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA03468
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 09:31:48 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3QDVTH51431;
	Fri, 26 Apr 2002 08:31:30 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CC956B1.FC0CC9DD@stewart.chicago.il.us>
Date: Fri, 26 Apr 2002 08:31:29 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: s_balani@trillium.com, qxie1@email.mot.com, mo_khan@trillium.com,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <58AEDE27D48F884285007ABF827FF26346A39E@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Ivan.Arias-Rodriguez@nokia.com wrote:
> 
>         Hi Sandeep!
> 
>         See my comment inline...
> 
> > hi all,
> >
> > My comments inline,
> >
> >
> >       Randy, see my comment inline...
> >
> > > Ivan.Arias-Rodriguez@nokia.com wrote:
> > > >
> > > >         Hi all!
> > > >
> > > >         Well, now I realized that I had a somehow different
> > > view about what was a legal addition... thanks for your comments.
> > > >
> > > >         Now I have another doubt... Imagine that we have an
> > > endpoint E1 with addresses IPa, IPb and port 1. We have an
> > > association A1 that only uses IPa, whose peer is IPz and port
> > > 65535. Thus the association A1 can be represented as [(IPa;
> > > Port1), (IPz; Port 65535)]. What happens now if we receive an
> > > INIT from IPz/65535 directed to IPb/1?
> > >
> > > A question...
> > >
> > > Why did you not include IPb in the INIT or INIT-ACK that
> > setup A1. You
> > > are
> > > supposed to tell the peer of all your addresses that are
> > bound to your
> > > endpoint. I think this is the big flaw in your question..
> >
> >       I did, but later on I deleted it using the AddIP
> > extension... :-) So
> > now I don't use it anymore...
> >
> > [ sandeep ] I think this is very interesting proble :-)
> >  What i think we can do is When u are doing DELETEIP for IPb, If no
> > association is using IPb as source address we should remover IPb from
> > Endpoint E1. This way we will not have any such problem.
> >
> > What do u think ????
> 
>         No, I don't think this is a good solution... Remember that there can be many associations in that endpoint, and that they might be using IPb as A1 did once... I don't think you can erase that address from the endpoint...

Hmm.. maybe the real solution is you MUST NOT use the DELETE unless
you are also deleting it from the endpoint.

So... if you do a DELETE IP-B you MUST first delete it from the
endpoint. Then you MAY delete it from any association... ones you
don't delete it from will have to have a variant list of addresses
that it has ...

That way everything still works for those that DO NOT delete IP-B and
your problem goes away...

Of course your problem STILL exists when you HOT-PLUG in a card
and have an existing association and NO ADD-IP extension...


R
-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Apr 26 10:03:50 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA04746
	for <tsvwg-archive@odin.ietf.org>; Fri, 26 Apr 2002 10:03:50 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA13857
	for tsvwg-archive@odin.ietf.org; Fri, 26 Apr 2002 10:03:53 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA12061;
	Fri, 26 Apr 2002 09:34:13 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA12018
	for <tsvwg@optimus.ietf.org>; Fri, 26 Apr 2002 09:34:09 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA03540
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 09:34:05 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3QDXvH51440;
	Fri, 26 Apr 2002 08:33:57 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CC95744.35D6E431@stewart.chicago.il.us>
Date: Fri, 26 Apr 2002 08:33:56 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: Dave.Sclarsky@radisys.com, tsvwg@ietf.org, Simon.Zhuang@radisys.com,
        Heinz.Prantner@radisys.com, Gene.Litt@radisys.com,
        Seamus.Gilchrist@radisys.com
Subject: Re: [Tsvwg] T3 cannot timeout???
References: <58AEDE27D48F884285007ABF827FF26308EFB0@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Ivan.Arias-Rodriguez@nokia.com wrote:
> 
>         Hi Dave!
> 
>         I must confess that I don't really follow you, but let's try...
> 
> > Hello,
> > We've found a Here's a case we found, hopefully someone can help!
> >
> > 2 peers (A and B) with 2 destinations (1 and 2) each, both
> > using the same
> > path as the 'primary'.
> 
>         Let me check if I know what you are saying here... Peer A uses IPa1 and IPa2, and peer B uses IPb1 and IPb2... So, you mean that A's primary address is IPb1 and uses IPa1 as the source address, and peer B uses IPa1 as primary and IPb1 as the source address for those datagrams, right?
> 
> > Primary ethernet goes down.
> 
>         Which one? IPb1?
> 
> > Side A detects this, and switches to destination 2.
> 
>         Ok, now A is using IPb2...
> 
> > Side B's T3 timer hasn't yet expired, but starts receiving
> > data from A on
> 
>         Ok... I think I understand you now that I read it again... You mean that there are two independent Ethernet networks... So IPa1 and IPb1 are in the same network, and IPa2 and IPb2 are in another network. Am I right?
> 
> > 2.  B sends SACK on 2nd path (since that's where it got the data), and
> > since it also has data to send, it bundles DATA chunks with the SACK.
> 
>         I don't think this is the right thing to do... You always use the primary address until it is down... I would most probably do it in the other way: sending SACKs bundled with the DATA chunks sent to the primary address...
> 

I agree with Ivan here.. You BUNDLE and send the SACK to
the PRIMARY address when you are putting it with DATA. NOT to
where it came from...


R


> > The data bundled with the (many?) SACKs on dest 2 will eventually be
> > SACKed.  These SACKs must contain GAP reports, indicating the
> > missing TSNs
> > (the ones sent to dest 1 with TSNs lower than those bundled with the
> > SACKs).  Once 4 of these SACKs with GAPs are received, Side B
> > will "fast
> > retransmit" those chunks on dest 2.  According to the RFC
> 
>         Well, that's right... However I think you shouldn't send the DATA to IPa2 just because you are also sending SACKs there...
> 
> > (7.2.4), Side B
> > will restart T3 for dest 1, since it will be retransmitting
> > the lowest TSN
> > sent to it.  It appears that this will continue forever.
> > Side B will never
> > see dest 1 as failed, since its T3 will continually be reset,
> > but it will
> > waste a lot of time and energy sending it data!
> 
>         Ok, so the problem is that you are sending DATA chunks both to IPa1 and IPa2 depending on if you send SACKs or not... I think this is the problem... T3 in IPa1 won't reset because the SACKs of the DATA chunks sent to IPa2 will always trigger a fast retransmission before it expires...
> 
>         However, if you use only the primary address when sending data, you don't have this problem... I don't know if this is the solution, but the point is that what you are doing sending DATA to the secondary address is not the right thing to do...
> 
>         BR Iván Arias Rodríguez
> 
> > Has anyone seen this?  Any suggestions?
> >
> > ======================================
> > Dave Sclarsky
> > Radisys Corporation
> > 520 Fellowship Road
> > Suite C-304
> > Mt. Laurel, NJ 08054
> > email:dave.sclarsky@radisys.com
> > voice:856.642.9229x231 fax:856.642.0252
> >
> >
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> >
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Apr 26 10:31:47 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA05893
	for <tsvwg-archive@odin.ietf.org>; Fri, 26 Apr 2002 10:31:47 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA15989
	for tsvwg-archive@odin.ietf.org; Fri, 26 Apr 2002 10:31:50 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA14498;
	Fri, 26 Apr 2002 10:11:03 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA14465
	for <tsvwg@optimus.ietf.org>; Fri, 26 Apr 2002 10:10:59 -0400 (EDT)
Received: from prue.eim.surrey.ac.uk (IDENT:exim@prue.eim.surrey.ac.uk [131.227.76.5])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA05022
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 10:10:51 -0400 (EDT)
Received: from artemis.ee.surrey.ac.uk ([131.227.88.18] ident=eep1lw)
	by prue.eim.surrey.ac.uk with esmtp (Exim 3.33 #4)
	id 1716Qo-0003nR-00; Fri, 26 Apr 2002 15:10:38 +0100
Date: Fri, 26 Apr 2002 15:10:37 +0100 (BST)
From: Lloyd Wood <l.wood@eim.surrey.ac.uk>
X-X-Sender: eep1lw@artemis.ee.surrey.ac.uk
Reply-To: Lloyd Wood <L.Wood@eim.surrey.ac.uk>
To: Dohy HONG <Dohy.Hong@ens.fr>
cc: end2end-interest@postel.org, <tsvwg@ietf.org>
In-Reply-To: <Pine.BSF.4.21.0204261347320.18699-100000@baobab.ens.fr>
Message-ID: <Pine.SOL.4.43.0204261505370.1987-100000@artemis.ee.surrey.ac.uk>
Organization: speaking for none
X-url: http://www.ee.surrey.ac.uk/Personal/L.Wood/
X-no-archive: yes
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=ISO-8859-1
X-Scanner: exiscan *1716Qo-0003nR-00*WR1tECZE8dQ* (SECM, UniS)
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from QUOTED-PRINTABLE to 8bit by optimus.ietf.org id KAA14466
Subject: [Tsvwg] Re: [e2e] FTCP algorithm
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

On Fri, 26 Apr 2002, Dohy HONG wrote separately to
tsvwg and end2end:

> I would like to open discussions about the ideas of
> FTCP algorithm presented in:
> http://www.ietf.org/internet-drafts/draft-hong-ftcp-00.txt

(sounds like a TCP rate-based pacing variant to me.)

opening up separate discussions on two lists seems unproductive;
I think end2end would be more appropriate for any followups,
since it's algorithm rather than implementation.

L.

> This algorithm propose a simple modification of
> TCP algorithm to improve the performance in a way
> which is 'socially better'.
> And it seems that there are 2 interesting properties
> in using FTCP (F for fluid):
> One is that it improve the End-to-End performance for
> its user.
> The second is that it encourages others to also use
> this algorithm and become 'socially better'.
>
> The basic idea is based on the remark that the
> packet burstiness inside each RTT is not good:
> especially with Slow Start and long RTTs.
> If you have CWND=10 with RTT~1 second, it is better
> to send 1 packet every 100 ms that to send a burst
> of 10 packets.
>
> What is your gain? I did some simple test modifying
> the linux kernel and doing FTP and HTTP
> (I also did some simple NS simulation which is described
> in the draft).
> My feeling was that you could gain about 0-50%
> depending on RTT, congestion state etc. It would be
> nice if other people could check this
> independently. I would be more than happy to help
> anyone who would like to do kernel modification
> (or for ns-2) for FTCP and get any feedback.
>
> What is the cost? Well, I think FTCP idea is more
> efficient with a good timer granularity.
> Timer is needed by the fact that you need to be
> patient before sending packets after ACKs reception
> (roughly speaking: try to send 1 packet every
> CWND/RTT).
> A good point is also that this idea could be
> implemented at the source or/and the destination level.
>
> I naively hope that FTCP could become useful to
> people in practice.
>
> All comments, criticisms (even blames) are very
> welcome.
> Dohy
>
> ===================================================
> Dohy HONG
> INRIA-ENS
> Projet TREC (Théorie des Réseaux Et Communications)
>
> ECOLE NORMALE SUPERIEURE
> Département Informatique
> 45 rue d'Ulm, 75230 PARIS CEDEX 05,  FRANCE
> Tel : 33 (1) 4432 2112, Fax : 33 (1) 4432 2151
>         ------------------------------------
> Email : Dohy.Hong@ens.fr
> http://www.di.ens.fr/~trec
> ===================================================

<L.Wood@surrey.ac.uk>PGP<http://www.ee.surrey.ac.uk/Personal/L.Wood/>




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Apr 26 11:45:32 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA08279
	for <tsvwg-archive@odin.ietf.org>; Fri, 26 Apr 2002 11:45:31 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA21057
	for tsvwg-archive@odin.ietf.org; Fri, 26 Apr 2002 11:45:35 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA19053;
	Fri, 26 Apr 2002 11:12:10 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA19022
	for <tsvwg@optimus.ietf.org>; Fri, 26 Apr 2002 11:12:07 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-230.dallas.net [209.44.42.230])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA07229
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 11:12:01 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id KAA05574;
	Fri, 26 Apr 2002 10:11:43 -0500
Date: Fri, 26 Apr 2002 10:11:43 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Randall Stewart <randall@stewart.chicago.il.us>
Cc: Ivan.Arias-Rodriguez@nokia.com, Dave.Sclarsky@radisys.com, tsvwg@ietf.org,
        Simon.Zhuang@radisys.com, Heinz.Prantner@radisys.com,
        Gene.Litt@radisys.com, Seamus.Gilchrist@radisys.com
Subject: Re: [Tsvwg] T3 cannot timeout???
Message-ID: <20020426101143.D24386@openss7.org>
Reply-To: bidulock@openss7.org
References: <58AEDE27D48F884285007ABF827FF26308EFB0@esebe017.NOE.Nokia.com> <3CC95744.35D6E431@stewart.chicago.il.us>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <3CC95744.35D6E431@stewart.chicago.il.us>; from randall@stewart.chicago.il.us on Fri, Apr 26, 2002 at 08:33:56AM -0500
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Randall,

On Fri, 26 Apr 2002, Randall Stewart wrote:

> Ivan.Arias-Rodriguez@nokia.com wrote:
> > 
> >         Hi Dave!
> > 
> >         I must confess that I don't really follow you, but let's try...
> > 
> > > Hello,
> > > We've found a Here's a case we found, hopefully someone can help!
> > >
> > > 2 peers (A and B) with 2 destinations (1 and 2) each, both
> > > using the same
> > > path as the 'primary'.
> > 
> >         Let me check if I know what you are saying here... Peer A uses IPa1 and IPa2, and peer B uses IPb1 and IPb2... So, you mean that A's primary address is IPb1 and uses IPa1 as the source address, and peer B uses IPa1 as primary and IPb1 as the source address for those datagrams, right?
> > 
> > > Primary ethernet goes down.
> > 
> >         Which one? IPb1?
> > 
> > > Side A detects this, and switches to destination 2.
> > 
> >         Ok, now A is using IPb2...
> > 
> > > Side B's T3 timer hasn't yet expired, but starts receiving
> > > data from A on
> > 
> >         Ok... I think I understand you now that I read it again... You mean that there are two independent Ethernet networks... So IPa1 and IPb1 are in the same network, and IPa2 and IPb2 are in another network. Am I right?
> > 
> > > 2.  B sends SACK on 2nd path (since that's where it got the data), and
> > > since it also has data to send, it bundles DATA chunks with the SACK.
> > 
> >         I don't think this is the right thing to do... You always use the primary address until it is down... I would most probably do it in the other way: sending SACKs bundled with the DATA chunks sent to the primary address...
> > 
> 
> I agree with Ivan here.. You BUNDLE and send the SACK to
> the PRIMARY address when you are putting it with DATA. NOT to
> where it came from...
> 
> 
> R

That's what I do too.  Works for me.

--brian



-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Apr 26 13:06:53 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA10897
	for <tsvwg-archive@odin.ietf.org>; Fri, 26 Apr 2002 13:06:53 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA26575
	for tsvwg-archive@odin.ietf.org; Fri, 26 Apr 2002 13:06:55 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA24968;
	Fri, 26 Apr 2002 12:39:23 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA24940
	for <tsvwg@ns.ietf.org>; Fri, 26 Apr 2002 12:39:20 -0400 (EDT)
Received: from hermes.fm.intel.com (fmr01.intel.com [192.55.52.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA10107
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 12:39:16 -0400 (EDT)
Received: from talaria.fm.intel.com (talaria.fm.intel.com [10.1.192.39])
	by hermes.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.41 2002/04/22 23:29:17 root Exp $) with ESMTP id g3QGeGM21015
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 16:40:16 GMT
Received: from fmsmsxvs042.fm.intel.com (fmsmsxv042-1.fm.intel.com [132.233.48.110])
	by talaria.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.16 2002/04/15 17:47:00 root Exp $) with SMTP id g3QGfNj22688
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 16:41:23 GMT
Received: from fmsmsx26.fm.intel.com ([132.233.42.26])
 by fmsmsxvs042.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002042609432328321
 ; Fri, 26 Apr 2002 09:43:24 -0700
Received: by fmsmsx26.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <2HL0Q27Y>; Fri, 26 Apr 2002 09:39:08 -0700
Message-ID: <F1CE15E08172D4119247009027AE9D50191173A5@FMSMSX37>
From: "Balani, Sandeep" <s_balani@trillium.com>
To: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        Ivan.Arias-Rodriguez@nokia.com
Cc: "Balani, Sandeep" <s_balani@trillium.com>, qxie1@email.mot.com,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Fri, 26 Apr 2002 09:39:05 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi All, 
  My comments inline... 

-sandeep  


-----Original Message-----
From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
Sent: Friday, April 26, 2002 6:31 AM
To: Ivan.Arias-Rodriguez@nokia.com
Cc: s_balani@trillium.com; qxie1@email.mot.com; mo_khan@trillium.com;
tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict


Ivan.Arias-Rodriguez@nokia.com wrote:
> 
>         Hi Sandeep!
> 
>         See my comment inline...
> 
> > hi all,
> >
> > My comments inline,
> >
> >
> >       Randy, see my comment inline...
> >
> > > Ivan.Arias-Rodriguez@nokia.com wrote:
> > > >
> > > >         Hi all!
> > > >
> > > >         Well, now I realized that I had a somehow different
> > > view about what was a legal addition... thanks for your comments.
> > > >
> > > >         Now I have another doubt... Imagine that we have an
> > > endpoint E1 with addresses IPa, IPb and port 1. We have an
> > > association A1 that only uses IPa, whose peer is IPz and port
> > > 65535. Thus the association A1 can be represented as [(IPa;
> > > Port1), (IPz; Port 65535)]. What happens now if we receive an
> > > INIT from IPz/65535 directed to IPb/1?
> > >
> > > A question...
> > >
> > > Why did you not include IPb in the INIT or INIT-ACK that
> > setup A1. You
> > > are
> > > supposed to tell the peer of all your addresses that are
> > bound to your
> > > endpoint. I think this is the big flaw in your question..
> >
> >       I did, but later on I deleted it using the AddIP
> > extension... :-) So
> > now I don't use it anymore...
> >
> > [ sandeep ] I think this is very interesting proble :-)
> >  What i think we can do is When u are doing DELETEIP for IPb, If no
> > association is using IPb as source address we should remover IPb from
> > Endpoint E1. This way we will not have any such problem.
> >
> > What do u think ????
> 
>         No, I don't think this is a good solution... Remember that there
can be >many associations in that endpoint, and that they might be using IPb
as A1 did >>once... I don't think you can erase that address from the
endpoint...

Hmm.. maybe the real solution is you MUST NOT use the DELETE unless
you are also deleting it from the endpoint.

So... if you do a DELETE IP-B you MUST first delete it from the
endpoint. Then you MAY delete it from any association... ones you
don't delete it from will have to have a variant list of addresses
that it has ...

That way everything still works for those that DO NOT delete IP-B and
your problem goes away...

Of course your problem STILL exists when you HOT-PLUG in a card
and have an existing association and NO ADD-IP extension...


[ Sandeep ]   Can we delete an IP address from an Endpoint first & then
delete It from Association. According  to the procedure, Between the time we
have send ASCONF( Delete IPb) & the time we receive ASCONF-ACK(Delete IPb)
we should accept all packets destined to IPb ????

Also deleting an IP address from an Endpoint means, Deleting it from
multiple associations simultaneously ( Or One-By-One ) , What if on One
association  this DeleteIP procedure fails ??? 

This may be prevented if We can put such recommendation which Randall
indicated in earlier mail, I.e. DELETEIP procedure is only recommended in
cases when NIC card assocaited with that Ip address is actually going down.
In that case we have to remove IP address from  Association & Endpoint .

I am really not sure whether the statment I made in previous mail should be
assumed TRUE/Correct  : "  An Association Can Use SUBSET of IP Addresses
from its Endpoint ". 

What do u guys think ??????  



R
-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Apr 26 13:09:15 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA11030
	for <tsvwg-archive@odin.ietf.org>; Fri, 26 Apr 2002 13:09:15 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA26801
	for tsvwg-archive@odin.ietf.org; Fri, 26 Apr 2002 13:09:17 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA25281;
	Fri, 26 Apr 2002 12:45:23 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA25251
	for <tsvwg@ns.ietf.org>; Fri, 26 Apr 2002 12:45:20 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA10321
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 12:45:17 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3QGixH51767;
	Fri, 26 Apr 2002 11:45:00 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CC9840B.FAC4749C@stewart.chicago.il.us>
Date: Fri, 26 Apr 2002 11:44:59 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: "Balani, Sandeep" <s_balani@trillium.com>
CC: Ivan.Arias-Rodriguez@nokia.com, qxie1@email.mot.com,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F1CE15E08172D4119247009027AE9D50191173A5@FMSMSX37>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

"Balani, Sandeep" wrote:
> 
> Hi All,
>   My comments inline...
> 
> -sandeep
> 
> -----Original Message-----
> From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: Friday, April 26, 2002 6:31 AM
> To: Ivan.Arias-Rodriguez@nokia.com
> Cc: s_balani@trillium.com; qxie1@email.mot.com; mo_khan@trillium.com;
> tsvwg@ietf.org
> Subject: Re: [Tsvwg] End point and association Address conflict
> 
> Ivan.Arias-Rodriguez@nokia.com wrote:
> >
> >         Hi Sandeep!
> >
> >         See my comment inline...
> >
> > > hi all,
> > >
> > > My comments inline,
> > >
> > >
> > >       Randy, see my comment inline...
> > >
> > > > Ivan.Arias-Rodriguez@nokia.com wrote:
> > > > >
> > > > >         Hi all!
> > > > >
> > > > >         Well, now I realized that I had a somehow different
> > > > view about what was a legal addition... thanks for your comments.
> > > > >
> > > > >         Now I have another doubt... Imagine that we have an
> > > > endpoint E1 with addresses IPa, IPb and port 1. We have an
> > > > association A1 that only uses IPa, whose peer is IPz and port
> > > > 65535. Thus the association A1 can be represented as [(IPa;
> > > > Port1), (IPz; Port 65535)]. What happens now if we receive an
> > > > INIT from IPz/65535 directed to IPb/1?
> > > >
> > > > A question...
> > > >
> > > > Why did you not include IPb in the INIT or INIT-ACK that
> > > setup A1. You
> > > > are
> > > > supposed to tell the peer of all your addresses that are
> > > bound to your
> > > > endpoint. I think this is the big flaw in your question..
> > >
> > >       I did, but later on I deleted it using the AddIP
> > > extension... :-) So
> > > now I don't use it anymore...
> > >
> > > [ sandeep ] I think this is very interesting proble :-)
> > >  What i think we can do is When u are doing DELETEIP for IPb, If no
> > > association is using IPb as source address we should remover IPb from
> > > Endpoint E1. This way we will not have any such problem.
> > >
> > > What do u think ????
> >
> >         No, I don't think this is a good solution... Remember that there
> can be >many associations in that endpoint, and that they might be using IPb
> as A1 did >>once... I don't think you can erase that address from the
> endpoint...
> 
> Hmm.. maybe the real solution is you MUST NOT use the DELETE unless
> you are also deleting it from the endpoint.
> 
> So... if you do a DELETE IP-B you MUST first delete it from the
> endpoint. Then you MAY delete it from any association... ones you
> don't delete it from will have to have a variant list of addresses
> that it has ...
> 
> That way everything still works for those that DO NOT delete IP-B and
> your problem goes away...
> 
> Of course your problem STILL exists when you HOT-PLUG in a card
> and have an existing association and NO ADD-IP extension...
> 
> [ Sandeep ]   Can we delete an IP address from an Endpoint first & then
> delete It from Association. According  to the procedure, Between the time we
> have send ASCONF( Delete IPb) & the time we receive ASCONF-ACK(Delete IPb)
> we should accept all packets destined to IPb ????
> 

I would think it would be ok.. remember the Endpoint is only
used for basic things like new assocaitions... the actual
"what addresses are in the association" can be kept in association
specific instances...


> Also deleting an IP address from an Endpoint means, Deleting it from
> multiple associations simultaneously ( Or One-By-One ) , What if on One
> association  this DeleteIP procedure fails ???

Well.. I don't see an issue.. so you are stuck with a deleted IP address
in your assocaition. 

A) Chances are you won't receive packets on it (since it is off the
   machine).

B) If it is NOT off the machine your code STILL should work in that
   it had better be keeping a list of addresses on the association that
   are valid... (when there is a deviant case such as this).

> 
> This may be prevented if We can put such recommendation which Randall
> indicated in earlier mail, I.e. DELETEIP procedure is only recommended in
> cases when NIC card assocaited with that Ip address is actually going down.
> In that case we have to remove IP address from  Association & Endpoint .

Yes.. but you still can end up with a IP address stuck in your
association if the peer does not do ADD-IP.. 

> 
> I am really not sure whether the statment I made in previous mail should be
> assumed TRUE/Correct  : "  An Association Can Use SUBSET of IP Addresses
> from its Endpoint ".
> 
> What do u guys think ??????

I think you CAN do this but it WILL cause you grief.. if by using
a subset you are saying making a decision that I am NOT going
to include IP-C in my address list (even though it is reachable
and should have normally been included).

R

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Apr 26 13:29:48 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA13931
	for <tsvwg-archive@odin.ietf.org>; Fri, 26 Apr 2002 13:29:48 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA08026
	for tsvwg-archive@odin.ietf.org; Fri, 26 Apr 2002 13:29:50 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA26655;
	Fri, 26 Apr 2002 13:07:21 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA26624
	for <tsvwg@ns.ietf.org>; Fri, 26 Apr 2002 13:07:17 -0400 (EDT)
Received: from caduceus.fm.intel.com (fmr02.intel.com [192.55.52.25])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA10966
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 13:07:12 -0400 (EDT)
Received: from talaria.fm.intel.com (talaria.fm.intel.com [10.1.192.39])
	by caduceus.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.41 2002/04/22 23:29:17 root Exp $) with ESMTP id g3QH6IT07698
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 17:06:18 GMT
Received: from fmsmsxvs042.fm.intel.com (fmsmsxv042-1.fm.intel.com [132.233.48.110])
	by talaria.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.16 2002/04/15 17:47:00 root Exp $) with SMTP id g3QH9JU20720
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 17:09:19 GMT
Received: from FMSMSX018.fm.intel.com ([132.233.42.197])
 by fmsmsxvs042.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002042610111312263
 ; Fri, 26 Apr 2002 10:11:14 -0700
Received: by fmsmsx018.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <2WWTMZLX>; Fri, 26 Apr 2002 10:06:59 -0700
Message-ID: <F1CE15E08172D4119247009027AE9D50191173A6@FMSMSX37>
From: "Balani, Sandeep" <s_balani@trillium.com>
To: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        "Balani, Sandeep" <s_balani@trillium.com>
Cc: Ivan.Arias-Rodriguez@nokia.com, qxie1@email.mot.com,
        "Khan, Mohammad"<mo_khan@trillium.com>, tsvwg@ietf.org
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Fri, 26 Apr 2002 10:06:48 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

my comments..

-----Original Message-----
From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
Sent: Friday, April 26, 2002 9:45 AM
To: Balani, Sandeep
Cc: Ivan.Arias-Rodriguez@nokia.com; qxie1@email.mot.com; Khan, Mohammad;
tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict


"Balani, Sandeep" wrote:
> 
> Hi All,
>   My comments inline...
> 
> -sandeep
> 
> -----Original Message-----
> From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: Friday, April 26, 2002 6:31 AM
> To: Ivan.Arias-Rodriguez@nokia.com
> Cc: s_balani@trillium.com; qxie1@email.mot.com; mo_khan@trillium.com;
> tsvwg@ietf.org
> Subject: Re: [Tsvwg] End point and association Address conflict
> 
> Ivan.Arias-Rodriguez@nokia.com wrote:
> >
> >         Hi Sandeep!
> >
> >         See my comment inline...
> >
> > > hi all,
> > >
> > > My comments inline,
> > >
> > >
> > >       Randy, see my comment inline...
> > >
> > > > Ivan.Arias-Rodriguez@nokia.com wrote:
> > > > >
> > > > >         Hi all!
> > > > >
> > > > >         Well, now I realized that I had a somehow different
> > > > view about what was a legal addition... thanks for your comments.
> > > > >
> > > > >         Now I have another doubt... Imagine that we have an
> > > > endpoint E1 with addresses IPa, IPb and port 1. We have an
> > > > association A1 that only uses IPa, whose peer is IPz and port
> > > > 65535. Thus the association A1 can be represented as [(IPa;
> > > > Port1), (IPz; Port 65535)]. What happens now if we receive an
> > > > INIT from IPz/65535 directed to IPb/1?
> > > >
> > > > A question...
> > > >
> > > > Why did you not include IPb in the INIT or INIT-ACK that
> > > setup A1. You
> > > > are
> > > > supposed to tell the peer of all your addresses that are
> > > bound to your
> > > > endpoint. I think this is the big flaw in your question..
> > >
> > >       I did, but later on I deleted it using the AddIP
> > > extension... :-) So
> > > now I don't use it anymore...
> > >
> > > [ sandeep ] I think this is very interesting proble :-)
> > >  What i think we can do is When u are doing DELETEIP for IPb, If no
> > > association is using IPb as source address we should remover IPb from
> > > Endpoint E1. This way we will not have any such problem.
> > >
> > > What do u think ????
> >
> >         No, I don't think this is a good solution... Remember that there
> can be >many associations in that endpoint, and that they might be using
IPb
> as A1 did >>once... I don't think you can erase that address from the
> endpoint...
> 
> Hmm.. maybe the real solution is you MUST NOT use the DELETE unless
> you are also deleting it from the endpoint.
> 
> So... if you do a DELETE IP-B you MUST first delete it from the
> endpoint. Then you MAY delete it from any association... ones you
> don't delete it from will have to have a variant list of addresses
> that it has ...
> 
> That way everything still works for those that DO NOT delete IP-B and
> your problem goes away...
> 
> Of course your problem STILL exists when you HOT-PLUG in a card
> and have an existing association and NO ADD-IP extension...
> 
> [ Sandeep ]   Can we delete an IP address from an Endpoint first & then
> delete It from Association. According  to the procedure, Between the time
we
> have send ASCONF( Delete IPb) & the time we receive ASCONF-ACK(Delete IPb)
> we should accept all packets destined to IPb ????
> 

I would think it would be ok.. remember the Endpoint is only
used for basic things like new assocaitions... the actual
"what addresses are in the association" can be kept in association
specific instances...

[ Sandeep ] I probably don't understand this point.  But i don't agree with
this 100%.  
This may be very implemenataion specific but For all incoming DATAGRAM i
should check if the Endpoint exist or Not, If Endpoint does not exist i will
not proceed .. So in case i deleted the IPb Address from the Endpoint, I
will never be able to accept the packet for that association which has IPb
in its source address. 

Moreover i think that its illegal to have a Source Address not present in
Endpoint & present in Association ( if that's the case ) ???  


> Also deleting an IP address from an Endpoint means, Deleting it from
> multiple associations simultaneously ( Or One-By-One ) , What if on One
> association  this DeleteIP procedure fails ???

Well.. I don't see an issue.. so you are stuck with a deleted IP address
in your assocaition. 

A) Chances are you won't receive packets on it (since it is off the
   machine).

B) If it is NOT off the machine your code STILL should work in that
   it had better be keeping a list of addresses on the association that
   are valid... (when there is a deviant case such as this).


> 
> This may be prevented if We can put such recommendation which Randall
> indicated in earlier mail, I.e. DELETEIP procedure is only recommended in
> cases when NIC card assocaited with that Ip address is actually going
down.
> In that case we have to remove IP address from  Association & Endpoint .

Yes.. but you still can end up with a IP address stuck in your
association if the peer does not do ADD-IP.. 



> 
> I am really not sure whether the statment I made in previous mail should
be
> assumed TRUE/Correct  : "  An Association Can Use SUBSET of IP Addresses
> from its Endpoint ".
> 
> What do u guys think ??????

I think you CAN do this but it WILL cause you grief.. if by using
a subset you are saying making a decision that I am NOT going
to include IP-C in my address list (even though it is reachable
and should have normally been included).


R

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Apr 26 14:43:29 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA23867
	for <tsvwg-archive@odin.ietf.org>; Fri, 26 Apr 2002 14:43:28 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA15626
	for tsvwg-archive@odin.ietf.org; Fri, 26 Apr 2002 14:43:30 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA13882;
	Fri, 26 Apr 2002 14:15:18 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA13854
	for <tsvwg@ns.ietf.org>; Fri, 26 Apr 2002 14:15:15 -0400 (EDT)
Received: from mail-gw.radisys.com (mail-gw.radisys.com [206.102.10.35])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA20393
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 14:15:11 -0400 (EDT)
From: Dave.Sclarsky@radisys.com
Received: by mail-gw.radisys.com (Postfix, from userid 5)
	id 1272C4E8AF; Fri, 26 Apr 2002 11:15:10 -0700 (PDT)
Received: from notes.radisys.com(206.103.52.220)
 via SMTP by mail-gw.radisys.com, id smtpdAAAg9ayHL; Fri Apr 26 11:15:02 2002
To: tsvwg@ietf.org
Cc: Gene.Litt@radisys.com, Heinz.Prantner@radisys.com,
        Ivan.Arias-Rodriguez@nokia.com,
        Randall Stewart <randall@stewart.chicago.il.us>,
        Seamus.Gilchrist@radisys.com, Simon.Zhuang@radisys.com,
        bidulock@openss7.org
Subject: Re: [Tsvwg] T3 cannot timeout???
MIME-Version: 1.0
X-Mailer: Lotus Notes Release 5.0.8  June 18, 2001
Message-ID: <OFEBE3E319.CF67A6BC-ON85256BA7.00620AEC@radisys.com>
Date: Fri, 26 Apr 2002 14:01:35 -0400
X-MIMETrack: Serialize by Router on HQ_ACS_1/Radisys_Corporation/US(Release 5.0.9a |January
 7, 2002) at 04/26/2002 11:06:50 AM,
	Serialize complete at 04/26/2002 11:06:50 AM
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id OAA13855
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

All,
Thanks for the prompt replies.  We tried your suggestion, and had somewhat 
better success, but....
OK, so we are now sending SACKs to the PRIMARY whenever they are bundled 
with DATA.  Well, when the PRIMARY cable is pulled, both sides start 
RTXing to dest 2. But, since we have heavy traffic, the SACKs for the 
RTX'd data is bundled with new data, and sent over the (dysfunctional) 
PRIMARY.  Whoops, the RTX'd data doesn't get ACK'd until each side 
declares the PRIMARY dead and switches to the 2nd dest.
I'm going to try instead to bundle SACKs with DATA ONLY if the 
destinations are the same, so that data RTX'd on non-primaries will have 
SACKs sent on the same non-primary, and we won't be able to bundle in this 
case.
Any thoughts on this?

DaveS.





"Brian F. G. Bidulock" <bidulock@openss7.org>
04/26/2002 11:11 AM
Please respond to bidulock

 
        To:     Randall Stewart <randall@stewart.chicago.il.us>
        cc:     Ivan.Arias-Rodriguez@nokia.com, Dave.Sclarsky@radisys.com, tsvwg@ietf.org, 
Simon.Zhuang@radisys.com, Heinz.Prantner@radisys.com, 
Gene.Litt@radisys.com, Seamus.Gilchrist@radisys.com
        Subject:        Re: [Tsvwg] T3 cannot timeout???


Randall,

On Fri, 26 Apr 2002, Randall Stewart wrote:

> Ivan.Arias-Rodriguez@nokia.com wrote:
> > 
> >         Hi Dave!
> > 
> >         I must confess that I don't really follow you, but let's 
try...
> > 
> > > Hello,
> > > We've found a Here's a case we found, hopefully someone can help!
> > >
> > > 2 peers (A and B) with 2 destinations (1 and 2) each, both
> > > using the same
> > > path as the 'primary'.
> > 
> >         Let me check if I know what you are saying here... Peer A uses 
IPa1 and IPa2, and peer B uses IPb1 and IPb2... So, you mean that A's 
primary address is IPb1 and uses IPa1 as the source address, and peer B 
uses IPa1 as primary and IPb1 as the source address for those datagrams, 
right?
> > 
> > > Primary ethernet goes down.
> > 
> >         Which one? IPb1?
> > 
> > > Side A detects this, and switches to destination 2.
> > 
> >         Ok, now A is using IPb2...
> > 
> > > Side B's T3 timer hasn't yet expired, but starts receiving
> > > data from A on
> > 
> >         Ok... I think I understand you now that I read it again... You 
mean that there are two independent Ethernet networks... So IPa1 and IPb1 
are in the same network, and IPa2 and IPb2 are in another network. Am I 
right?
> > 
> > > 2.  B sends SACK on 2nd path (since that's where it got the data), 
and
> > > since it also has data to send, it bundles DATA chunks with the 
SACK.
> > 
> >         I don't think this is the right thing to do... You always use 
the primary address until it is down... I would most probably do it in the 
other way: sending SACKs bundled with the DATA chunks sent to the primary 
address...
> > 
> 
> I agree with Ivan here.. You BUNDLE and send the SACK to
> the PRIMARY address when you are putting it with DATA. NOT to
> where it came from...
> 
> 
> R

That's what I do too.  Works for me.

--brian



-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Apr 26 15:06:01 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA26739
	for <tsvwg-archive@odin.ietf.org>; Fri, 26 Apr 2002 15:05:56 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA17347
	for tsvwg-archive@odin.ietf.org; Fri, 26 Apr 2002 15:05:59 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA15219;
	Fri, 26 Apr 2002 14:36:54 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA15190
	for <tsvwg@ns.ietf.org>; Fri, 26 Apr 2002 14:36:51 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA23190
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 14:36:47 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3QIahH51940;
	Fri, 26 Apr 2002 13:36:44 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CC99E3B.D3A4766A@stewart.chicago.il.us>
Date: Fri, 26 Apr 2002 13:36:43 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Dave.Sclarsky@radisys.com
CC: tsvwg@ietf.org, Gene.Litt@radisys.com, Heinz.Prantner@radisys.com,
        Ivan.Arias-Rodriguez@nokia.com, Seamus.Gilchrist@radisys.com,
        Simon.Zhuang@radisys.com, bidulock@openss7.org
Subject: Re: [Tsvwg] T3 cannot timeout???
References: <OFEBE3E319.CF67A6BC-ON85256BA7.00620AEC@radisys.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Dave:

Another little trick that you may not be doing 
when you recieve duplicate data chunks .. you retransmit to
alternate destinations for the SACK. 

This in some cases means you WON'T bundle the SACK's and the
DATA.. but thats ok.

R

Dave.Sclarsky@radisys.com wrote:
> 
> All,
> Thanks for the prompt replies.  We tried your suggestion, and had somewhat
> better success, but....
> OK, so we are now sending SACKs to the PRIMARY whenever they are bundled
> with DATA.  Well, when the PRIMARY cable is pulled, both sides start
> RTXing to dest 2. But, since we have heavy traffic, the SACKs for the
> RTX'd data is bundled with new data, and sent over the (dysfunctional)
> PRIMARY.  Whoops, the RTX'd data doesn't get ACK'd until each side
> declares the PRIMARY dead and switches to the 2nd dest.
> I'm going to try instead to bundle SACKs with DATA ONLY if the
> destinations are the same, so that data RTX'd on non-primaries will have
> SACKs sent on the same non-primary, and we won't be able to bundle in this
> case.
> Any thoughts on this?
> 
> DaveS.
> 
> "Brian F. G. Bidulock" <bidulock@openss7.org>
> 04/26/2002 11:11 AM
> Please respond to bidulock
> 
> 
>         To:     Randall Stewart <randall@stewart.chicago.il.us>
>         cc:     Ivan.Arias-Rodriguez@nokia.com, Dave.Sclarsky@radisys.com, tsvwg@ietf.org,
> Simon.Zhuang@radisys.com, Heinz.Prantner@radisys.com,
> Gene.Litt@radisys.com, Seamus.Gilchrist@radisys.com
>         Subject:        Re: [Tsvwg] T3 cannot timeout???
> 
> Randall,
> 
> On Fri, 26 Apr 2002, Randall Stewart wrote:
> 
> > Ivan.Arias-Rodriguez@nokia.com wrote:
> > >
> > >         Hi Dave!
> > >
> > >         I must confess that I don't really follow you, but let's
> try...
> > >
> > > > Hello,
> > > > We've found a Here's a case we found, hopefully someone can help!
> > > >
> > > > 2 peers (A and B) with 2 destinations (1 and 2) each, both
> > > > using the same
> > > > path as the 'primary'.
> > >
> > >         Let me check if I know what you are saying here... Peer A uses
> IPa1 and IPa2, and peer B uses IPb1 and IPb2... So, you mean that A's
> primary address is IPb1 and uses IPa1 as the source address, and peer B
> uses IPa1 as primary and IPb1 as the source address for those datagrams,
> right?
> > >
> > > > Primary ethernet goes down.
> > >
> > >         Which one? IPb1?
> > >
> > > > Side A detects this, and switches to destination 2.
> > >
> > >         Ok, now A is using IPb2...
> > >
> > > > Side B's T3 timer hasn't yet expired, but starts receiving
> > > > data from A on
> > >
> > >         Ok... I think I understand you now that I read it again... You
> mean that there are two independent Ethernet networks... So IPa1 and IPb1
> are in the same network, and IPa2 and IPb2 are in another network. Am I
> right?
> > >
> > > > 2.  B sends SACK on 2nd path (since that's where it got the data),
> and
> > > > since it also has data to send, it bundles DATA chunks with the
> SACK.
> > >
> > >         I don't think this is the right thing to do... You always use
> the primary address until it is down... I would most probably do it in the
> other way: sending SACKs bundled with the DATA chunks sent to the primary
> address...
> > >
> >
> > I agree with Ivan here.. You BUNDLE and send the SACK to
> > the PRIMARY address when you are putting it with DATA. NOT to
> > where it came from...
> >
> >
> > R
> 
> That's what I do too.  Works for me.
> 
> --brian
> 
> --
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Apr 26 15:28:14 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA29524
	for <tsvwg-archive@odin.ietf.org>; Fri, 26 Apr 2002 15:28:09 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA18226
	for tsvwg-archive@odin.ietf.org; Fri, 26 Apr 2002 15:28:12 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA17266;
	Fri, 26 Apr 2002 15:03:48 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA17237
	for <tsvwg@ns.ietf.org>; Fri, 26 Apr 2002 15:03:45 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-230.dallas.net [209.44.42.230])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA26469
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 15:03:39 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id OAA08629;
	Fri, 26 Apr 2002 14:03:25 -0500
Date: Fri, 26 Apr 2002 14:03:25 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Randall Stewart <randall@stewart.chicago.il.us>
Cc: Dave.Sclarsky@radisys.com, tsvwg@ietf.org, Gene.Litt@radisys.com,
        Heinz.Prantner@radisys.com, Ivan.Arias-Rodriguez@nokia.com,
        Seamus.Gilchrist@radisys.com, Simon.Zhuang@radisys.com
Subject: Re: [Tsvwg] T3 cannot timeout???
Message-ID: <20020426140325.L24386@openss7.org>
Reply-To: bidulock@openss7.org
References: <OFEBE3E319.CF67A6BC-ON85256BA7.00620AEC@radisys.com> <3CC99E3B.D3A4766A@stewart.chicago.il.us>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <3CC99E3B.D3A4766A@stewart.chicago.il.us>; from randall@stewart.chicago.il.us on Fri, Apr 26, 2002 at 01:36:43PM -0500
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Randall,

Again, that's what I do too. (And it is stated in the RFC.) The
first retrans in Dave's scenario results in dups which are used
as an indication that SACKs aren't being received and should be
sent to an alternate destination.

--brian

On Fri, 26 Apr 2002, Randall Stewart wrote:

> Dave:
> 
> Another little trick that you may not be doing 
> when you recieve duplicate data chunks .. you retransmit to
> alternate destinations for the SACK. 
> 
> This in some cases means you WON'T bundle the SACK's and the
> DATA.. but thats ok.
> 
> R
> 
> Dave.Sclarsky@radisys.com wrote:
> > 
> > All,
> > Thanks for the prompt replies.  We tried your suggestion, and had somewhat
> > better success, but....
> > OK, so we are now sending SACKs to the PRIMARY whenever they are bundled
> > with DATA.  Well, when the PRIMARY cable is pulled, both sides start
> > RTXing to dest 2. But, since we have heavy traffic, the SACKs for the
> > RTX'd data is bundled with new data, and sent over the (dysfunctional)
> > PRIMARY.  Whoops, the RTX'd data doesn't get ACK'd until each side
> > declares the PRIMARY dead and switches to the 2nd dest.
> > I'm going to try instead to bundle SACKs with DATA ONLY if the
> > destinations are the same, so that data RTX'd on non-primaries will have
> > SACKs sent on the same non-primary, and we won't be able to bundle in this
> > case.
> > Any thoughts on this?
> > 
> > DaveS.
> > 
> > "Brian F. G. Bidulock" <bidulock@openss7.org>
> > 04/26/2002 11:11 AM
> > Please respond to bidulock
> > 
> > 
> >         To:     Randall Stewart <randall@stewart.chicago.il.us>
> >         cc:     Ivan.Arias-Rodriguez@nokia.com, Dave.Sclarsky@radisys.com, tsvwg@ietf.org,
> > Simon.Zhuang@radisys.com, Heinz.Prantner@radisys.com,
> > Gene.Litt@radisys.com, Seamus.Gilchrist@radisys.com
> >         Subject:        Re: [Tsvwg] T3 cannot timeout???
> > 
> > Randall,
> > 
> > On Fri, 26 Apr 2002, Randall Stewart wrote:
> > 
> > > Ivan.Arias-Rodriguez@nokia.com wrote:
> > > >
> > > >         Hi Dave!
> > > >
> > > >         I must confess that I don't really follow you, but let's
> > try...
> > > >
> > > > > Hello,
> > > > > We've found a Here's a case we found, hopefully someone can help!
> > > > >
> > > > > 2 peers (A and B) with 2 destinations (1 and 2) each, both
> > > > > using the same
> > > > > path as the 'primary'.
> > > >
> > > >         Let me check if I know what you are saying here... Peer A uses
> > IPa1 and IPa2, and peer B uses IPb1 and IPb2... So, you mean that A's
> > primary address is IPb1 and uses IPa1 as the source address, and peer B
> > uses IPa1 as primary and IPb1 as the source address for those datagrams,
> > right?
> > > >
> > > > > Primary ethernet goes down.
> > > >
> > > >         Which one? IPb1?
> > > >
> > > > > Side A detects this, and switches to destination 2.
> > > >
> > > >         Ok, now A is using IPb2...
> > > >
> > > > > Side B's T3 timer hasn't yet expired, but starts receiving
> > > > > data from A on
> > > >
> > > >         Ok... I think I understand you now that I read it again... You
> > mean that there are two independent Ethernet networks... So IPa1 and IPb1
> > are in the same network, and IPa2 and IPb2 are in another network. Am I
> > right?
> > > >
> > > > > 2.  B sends SACK on 2nd path (since that's where it got the data),
> > and
> > > > > since it also has data to send, it bundles DATA chunks with the
> > SACK.
> > > >
> > > >         I don't think this is the right thing to do... You always use
> > the primary address until it is down... I would most probably do it in the
> > other way: sending SACKs bundled with the DATA chunks sent to the primary
> > address...
> > > >
> > >
> > > I agree with Ivan here.. You BUNDLE and send the SACK to
> > > the PRIMARY address when you are putting it with DATA. NOT to
> > > where it came from...
> > >
> > >
> > > R
> > 
> > That's what I do too.  Works for me.
> > 
> > --brian
> > 
> > --
> > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> >                         ¦ Therefore  all  progress  depends on the ¦
> >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Apr 26 17:57:57 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA17721
	for <tsvwg-archive@odin.ietf.org>; Fri, 26 Apr 2002 17:57:57 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA26677
	for tsvwg-archive@odin.ietf.org; Fri, 26 Apr 2002 17:58:01 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA24711;
	Fri, 26 Apr 2002 17:23:25 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA24611
	for <tsvwg@ns.ietf.org>; Fri, 26 Apr 2002 17:23:19 -0400 (EDT)
Received: from hermes.fm.intel.com (fmr01.intel.com [192.55.52.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA13300
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 17:23:12 -0400 (EDT)
Received: from talaria.fm.intel.com (talaria.fm.intel.com [10.1.192.39])
	by hermes.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.41 2002/04/22 23:29:17 root Exp $) with ESMTP id g3QLODs25087
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 21:24:13 GMT
Received: from fmsmsxvs043.fm.intel.com (fmsmsxv043-1.fm.intel.com [132.233.48.128])
	by talaria.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.16 2002/04/15 17:47:00 root Exp $) with SMTP id g3QLPP829975
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 21:25:25 GMT
Received: from fmsmsx019.fm.intel.com ([132.233.42.130])
 by fmsmsxvs043.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002042614221718504
 ; Fri, 26 Apr 2002 14:22:17 -0700
Received: by fmsmsx019.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <2XMHR1M9>; Fri, 26 Apr 2002 14:23:13 -0700
Message-ID: <F1CE15E08172D4119247009027AE9D50191173A8@FMSMSX37>
From: "Balani, Sandeep" <s_balani@trillium.com>
To: "'Qiaobing Xie'" <qxie1@email.mot.com>, Ivan.Arias-Rodriguez@nokia.com
Cc: randall@stewart.chicago.il.us, "Khan, Mohammad" <mo_khan@trillium.com>,
        tsvwg@ietf.org
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Fri, 26 Apr 2002 14:23:04 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org



> -----Original Message-----
> From: Qiaobing Xie [mailto:qxie1@email.mot.com]
> Sent: Friday, April 26, 2002 12:54 PM
> To: Ivan.Arias-Rodriguez@nokia.com
> Cc: Qiaobing Xie; randall@stewart.chicago.il.us; 
> s_balani@trillium.com;
> mo_khan@trillium.com; tsvwg@ietf.org
> Subject: Re: [Tsvwg] End point and association Address conflict
> 
> 
> Hi, Ivan,
> 
> This is a very important discussion. Please see my comments below...
> 
> Ivan.Arias-Rodriguez@nokia.com wrote:
> ....
> > > Here your "..include all the addr..." directly violated 
> the "under no
> > > circumstance can an SCTP transport address ever be present in
> > > more than
> > > one SCTP endpoint" rule (see last the paragraph, page 23 
> of the SCTP
> > > book).
> > >
> > >         E1               E2
> > > A1 = [(IPa; Port1), (IPz; Port 65535)]
> > >
> > >         E3                 E2
> > > A2 = [(IPa, IPb; Port1), (IPz; Port 65535)]
> > >
> > > And you have IPa/Port1 present in both E1 and E3 (they 
> are TWO different 
> > > endpoints by definition!). This is prohibited by the SCTP 
> association
> > > definition.
> > 
> >         No, I don't think what I'm proposing is illegal at 
> all... Take into account that you can make use of the AddIP 
> extension to delete IPb... Thus, initially, the association 
> had all the addresses of the endpoint (IPa and IPb). Later on 
> you delete your source address IPb. And then all this happens...
> 
> Applying Add/Delete-IP is ok, but anytime you do that, you must not
> break the basic rules that define the endpoint and association.
> Otherwise, you will get into troubles as you've noticed in your own
> analysis.
> 
> Let us layout what you are proposing in details.
> 
> Suppose you originally have:
> 
>           E1                  E2
>   A1 = [(IPa,IPb; Port1), (IPz; Port 65535)]
> 
> then you delete IPb from E1 (Note, this effectively changes the
> definition of E1 to E1' and A1 to A1'. They are logically differently
> defined now):
> 
>           E1'                  E2
>   A1' = [(IPa; Port1), (IPz; Port 65535)]


[ Sandeep ]  A slight change in the example above. What if initally I have 2
Associations. 

           E1                  E2
  A1 = [(IPa,IPb; Port1), (IPz; Port 65535)]
   
              E1                      E2
  A9 = [ (IPa, IPb; Port1), (IPm; Port 76577)]
 

And then i perform DELETEIP(IPb) from association A1,  then i have something
like 

          E1'                  E2
    A1' = [(IPa; Port1), (IPz; Port 65535)]

    And 
                 E1                      E2
    A9 = [ (IPa, IPb; Port1), (IPm; Port 76577)]


 Which is also illegal ....

  So what should be the option in this case for DELETEIP : 

(1) Delete IPb from the Endpoint & From Association A9 also. ( Assuming that
IPb  NIC Card is Down )  ??????????
(2) Create a Totally new Endpoint [ E1' ] & Delete E1 ?????? 

Or anything else ???????????????/


> 
> Now, E2 sends an INIT to IPb/Port1, and you decide it's ok to setup
> assoc A2, which is ok. But you are _PROBINITED_ by the endpoint
> definition to have:
> 
>           E1'                  E2
>   A1' = [(IPa; Port1), (IPz; Port 65535)]
> 
>           E3                  E2
>   A2 = [(IPa,IPb; Port1), (IPz; Port 65535)]  /* bad, illegal assoc */
> 
> I don't know how you can detect and prevent this illegal 
> setup conflict
> in your implementation. But allowing this will violate the basic SCTP
> association rule I mentioned above. Another way to understand why this
> is illegal is to think about this: if this were to be allowed, what
> would happen if you then go ahead and add IPb back to A1'? You would
> magically end up with _two_ association between the same pair of
> endpoints!?! 
> 
> However, it is perfectly ok to have:
> 
>           E1'                  E2
>   A1' = [(IPa; Port1), (IPz; Port 65535)]
> 
>           E3                  E2
>   A2 = [(IPb; Port1), (IPz; Port 65535)]  /* good, legal assoc */
> 
> This discussion is very good. We should capture this and 
> caution people
> not to break SCTP association definitions when using ADD/DELETE-IP.
> 
> -Qiaobing
> 
> > 
> >         So, the problem is that the pair IPb/1-IPz/65535 is 
> not part of any association... :-/
> > 
> >         BR Iván Arias Rodríguez
> > 
> > > -Qiaobing
> > >
> > > > in the INIT ACK chunk... But we have to be careful, because in
> > > > this case, when receiving the COOKIE ECHO, we have two options:
> > >
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Apr 26 19:00:03 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA24606
	for <tsvwg-archive@odin.ietf.org>; Fri, 26 Apr 2002 19:00:03 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id TAA29946
	for tsvwg-archive@odin.ietf.org; Fri, 26 Apr 2002 19:00:06 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA28510;
	Fri, 26 Apr 2002 18:35:22 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA28481
	for <tsvwg@ns.ietf.org>; Fri, 26 Apr 2002 18:35:19 -0400 (EDT)
Received: from hermes.fm.intel.com (fmr01.intel.com [192.55.52.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA21655
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 18:34:29 -0400 (EDT)
Received: from talaria.fm.intel.com (talaria.fm.intel.com [10.1.192.39])
	by hermes.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.41 2002/04/22 23:29:17 root Exp $) with ESMTP id g3QMZVi08073
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 22:35:31 GMT
Received: from fmsmsxvs040.fm.intel.com (fmsmsxv040-1.fm.intel.com [132.233.48.108])
	by talaria.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.16 2002/04/15 17:47:00 root Exp $) with SMTP id g3QMaiU24330
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 22:36:44 GMT
Received: from fmsmsx28.fm.intel.com ([132.233.42.28])
 by fmsmsxvs040.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002042615335405535
 ; Fri, 26 Apr 2002 15:33:55 -0700
Received: by fmsmsx28.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <2KDW2V03>; Fri, 26 Apr 2002 15:34:32 -0700
Message-ID: <F1CE15E08172D4119247009027AE9D50191173A9@FMSMSX37>
From: "Balani, Sandeep" <s_balani@trillium.com>
To: "'Qiaobing Xie'" <qxie1@email.mot.com>,
        "Balani, Sandeep" <s_balani@trillium.com>
Cc: "'Qiaobing Xie'" <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, randall@stewart.chicago.il.us,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Fri, 26 Apr 2002 15:34:27 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi Qiaobing, 
  My comments.

thanks
sandeep

> -----Original Message-----
> From: Qiaobing Xie [mailto:qxie1@email.mot.com]
> Sent: Friday, April 26, 2002 3:24 PM
> To: Balani, Sandeep
> Cc: 'Qiaobing Xie'; Ivan.Arias-Rodriguez@nokia.com;
> randall@stewart.chicago.il.us; Khan, Mohammad; tsvwg@ietf.org
> Subject: Re: [Tsvwg] End point and association Address conflict
> 
> 
> Hi, Sandeep,
> 
> It is very good point. Please see below for my answer.
> 
> > [ Sandeep ]  A slight change in the example above. What if 
> initally I have 2
> > Associations.
> > 
> >            E1                  E2
> >   A1 = [(IPa,IPb; Port1), (IPz; Port 65535)]
> > 
> >               E1                      E2
> >   A9 = [ (IPa, IPb; Port1), (IPm; Port 76577)]
> >
> > And then i perform DELETEIP(IPb) from association A1,  then 
> i have something
> > like
> > 
> >           E1'                  E2
> >     A1' = [(IPa; Port1), (IPz; Port 65535)]
> > 
> >     And
> >                  E1                      E2  /* call it E3 */
> >     A9 = [ (IPa, IPb; Port1), (IPm; Port 76577)]
> > 
> >  Which is also illegal ....
> 
> Yes, this is illegal. Your implementation should never allow this to
> happen. 
> 
> You may have some confusion over an important definition in 
> SCTP. It is
> conceptually WRONG to say "deleting IPb from association A1". One can
> only add or delete addresses to or from __an SCTP endpoint__. 
> 

[ Sandeep ] I totally totally agree with U on that " Addition/Deletion of IP
Address should be on an Endpoint & not on Association.  After all we are
modifying the source IP addresses which is an essential part of Endpoint " 

But Currently ADDIP/DELETEIP Draft Says : 

"


- Dynamic IP address reconfiguration extension: For
    computational or networking platforms that allow addition/removal of
    physical interface cards this feature can provide:

     A) a graceful method to add to the interfaces of an existing 
        *******association*******. For IPv6 this feature allows renumbering
        of existing associations.
"

There are similar references like this in the DRAFT, which mentions
ADD/DELETE IP procedures assocaitions-wise. 


So Is there any plans to change the defination of ADDIP/DELETEIP because If
i would do ADD/DELETE on Endpoint I have to perform ASCONF/ASCONF-ACK
procedures on all the assocations  belonging to that endpoint which seems
correct to me but In that case I won't be  in compliance with draft  ....


> With this important clarification of concept, all above should become
> very simple and clear - what you are allowed to do here is __to delete
> IPb from endpoint E1 which is currently being used in associations A1
> and A9__. If your implementation is doing the right thing, the result
> should be 
> 
>                E1'             E2
>     A1' = [(IPa; Port1), (IPz; Port 65535)]
> 
>     And
>                E1'                   E2  /* call it E3 */
>     A9' = [ (IPa; Port1), (IPm; Port 76577)]
> 
> If your intention is to transform the original A1 and A9 to the
> following:
> 
>                E1'             E2
>     A1' = [(IPa; Port1), (IPz; Port 65535)]
> 
>     And
>                E1'                   E2  /* call it E3 */
>     A9' = [ (IPb; Port1), (IPm; Port 76577)]
> 
> I am afraid that you have to tear down A1/A9 and set-up A1' and A9'
> fresh. ADD/DELET-IP will NOT get your here without violating the basic
> association rules.
> 
> -Qiaobing
> 

Thanks
Sandeep 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Apr 27 09:26:50 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA03933
	for <tsvwg-archive@odin.ietf.org>; Sat, 27 Apr 2002 09:26:49 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA10962
	for tsvwg-archive@odin.ietf.org; Sat, 27 Apr 2002 09:26:52 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA09855;
	Sat, 27 Apr 2002 08:59:44 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA09827
	for <tsvwg@optimus.ietf.org>; Sat, 27 Apr 2002 08:59:41 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA01208
	for <tsvwg@ietf.org>; Sat, 27 Apr 2002 08:59:36 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3RCx4H55741;
	Sat, 27 Apr 2002 07:59:04 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCAA097.7C1F6726@stewart.chicago.il.us>
Date: Sat, 27 Apr 2002 07:59:04 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Qiaobing Xie <qxie1@email.mot.com>
CC: "Balani, Sandeep" <s_balani@trillium.com>,
        "'Qiaobing Xie'" <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F1CE15E08172D4119247009027AE9D50191173A9@FMSMSX37> <3CC9E579.7DB84355@email.mot.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Qiabing/All:

I agree 100% with Q here. We need to tighten the wording up
but I don't think anything needs to be redefined... Now I
want to bring up something for all of you to consider... 

FORGET ADD-IP for a MINUTE...


In a NORMAL RFC2960 association I have the following

E1-(Bound INADDR_ANY)

I currently have

IP-A: Nic1/IP-B: Nic2

I setup an association with a peer

A1=(IP-A,IP-B:Port 4444)<====================>(IP-Z:Port 5555)

Now everything is fine...

Now My administrator plugs in Nic3 and does:

ifconfig nic3 IP-C

Now ... my endpoint has JUST CHANGED!! It is now

IP-A: Nic1/IP-B: Nic2/IP-C: Nic3

(remember we are bound to ALL)....

Now  my peer does 

                 <<------TO:IP-C Port4444----INIT(IP:Z port 5555)

.....

You have the SAME EXACT thing you have been dicussing ..


This is one of the reasons we NEED Add-IP. 

....If we think about the ADD-IP case... then when the ADDRESS
was added to the ENDPOINT (automatically) ... you would have gotten

-----ADD-IP(IP-C)-------->

and not reached the above condition...

The key is has Qiaobing stated, you ADD addresses to ENDPOINTs. I
think we will add some text towards this clearfication in the next
pass..

But the BIG thing I want to point our here is that this issue is
NOT because of ADD-IP but is something that can happen WITHOUT
ADD-IP... This is VERY important to remember... Add-ip actually
helps this situation... as long as your implementation of it
adds addresses to endpoints that is :->

R

Qiaobing Xie wrote:
> 
> Hi, Sandeep,
> 
> > [ Sandeep ] I totally totally agree with U on that " Addition/Deletion of IP
> > Address should be on an Endpoint & not on Association.  After all we are
> > modifying the source IP addresses which is an essential part of Endpoint "
> >
> > But Currently ADDIP/DELETEIP Draft Says :
> >
> > "
> > - Dynamic IP address reconfiguration extension: For
> >     computational or networking platforms that allow addition/removal of
> >     physical interface cards this feature can provide:
> >
> >      A) a graceful method to add to the interfaces of an existing
> >         *******association*******. For IPv6 this feature allows renumbering
> >         of existing associations.
> > "
> 
> I agree with you. These are loose wordings which may need to be
> tightened up in the draft. But in a general sense, with the correct
> understanding and clarification of the rules, the above wording is still
> fine - it simply depicts the final beneficiary of this add/delete-ip
> feature, which is indeed the associations themselves.
> 
> More importantly, in the draft we need to explicit notify the
> implementor that using add-ip/delete-ip is fine and benefiting, but one
> must not break the association rules, otherwise, it becomes a misuse of
> a good tool and you implementation will break.
> 
> I am glad this is brought up in this discussion thread.
> 
> >
> > There are similar references like this in the DRAFT, which mentions
> > ADD/DELETE IP procedures assocaitions-wise.
> 
> We will improve the text and make it tighter.
> 
> > So Is there any plans to change the defination of ADDIP/DELETEIP because If
> > i would do ADD/DELETE on Endpoint I have to perform ASCONF/ASCONF-ACK
> > procedures on all the assocations  belonging to that endpoint which seems
> > correct to me but In that case I won't be  in compliance with draft  ....
> 
> I don't think anything needs to be re-defined. We just need more
> clarifications. To the implementor, a good add/delete-ip API semantics
> might help to provide a safeguard against misuses.
> 
> SCTP is different from TCP in that it allows a single endpoint to
> associate with multiple peers (the "fan-out" config). But no matter how
> many peers the endpoint is concurrently connected to, logically it is
> _still_ a single endpoint.
> 
> With this being cleared, it should not be hard to understand that it is
> conceptually wrong to let a user apply add/delete-ip to only a subset of
> the fan-out concurrent connections on that single endpoint. Because this
> would in fact result in "splitting" the single endpoint into two
> different ones. This is NEVER the intention of add/delete-ip - the
> feature is meant to dynamically modify an endpoint without tearing down
> all the associations it currently has; it is NOT meant to split a single
> endpoint into two!
> 
> -Qiaobing
> 
> >
> > > With this important clarification of concept, all above should become
> > > very simple and clear - what you are allowed to do here is __to delete
> > > IPb from endpoint E1 which is currently being used in associations A1
> > > and A9__. If your implementation is doing the right thing, the result
> > > should be
> > >
> > >                E1'             E2
> > >     A1' = [(IPa; Port1), (IPz; Port 65535)]
> > >
> > >     And
> > >                E1'                   E2  /* call it E3 */
> > >     A9' = [ (IPa; Port1), (IPm; Port 76577)]
> > >
> > > If your intention is to transform the original A1 and A9 to the
> > > following:
> > >
> > >                E1'             E2
> > >     A1' = [(IPa; Port1), (IPz; Port 65535)]
> > >
> > >     And
> > >                E1'                   E2  /* call it E3 */
> > >     A9' = [ (IPb; Port1), (IPm; Port 76577)]
> > >
> > > I am afraid that you have to tear down A1/A9 and set-up A1' and A9'
> > > fresh. ADD/DELET-IP will NOT get your here without violating the basic
> > > association rules.
> > >
> > > -Qiaobing
> > >
> >
> > Thanks
> > Sandeep

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Apr 27 17:13:34 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA24454
	for <tsvwg-archive@odin.ietf.org>; Sat, 27 Apr 2002 17:13:34 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA28977
	for tsvwg-archive@odin.ietf.org; Sat, 27 Apr 2002 17:13:38 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA26700;
	Sat, 27 Apr 2002 16:44:11 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA25193
	for <tsvwg@ns.ietf.org>; Tue, 23 Apr 2002 15:45:25 -0400 (EDT)
Received: from motgate.mot.com (motgate.mot.com [129.188.136.100])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA04248
	for <tsvwg@ietf.org>; Tue, 23 Apr 2002 15:45:20 -0400 (EDT)
Received: [from pobox4.mot.com (pobox4.mot.com [10.64.251.243]) by motgate.mot.com (motgate 2.1) with ESMTP id MAA12780; Tue, 23 Apr 2002 12:45:23 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox4.mot.com (MOT-pobox4 2.0) with ESMTP id MAA22188; Tue, 23 Apr 2002 12:45:22 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3NJfej13680;
	Tue, 23 Apr 2002 14:41:40 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id OAA23746;
	Tue, 23 Apr 2002 14:41:47 -0500 (CDT)
Message-ID: <3CC5B93C.7E71EBFE@email.mot.com>
Date: Tue, 23 Apr 2002 14:42:52 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: "Balani, Sandeep" <s_balani@trillium.com>
CC: "'Qiaobing Xie'" <Qiaobing.Xie@motorola.com>,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F1CE15E08172D4119247009027AE9D5019117392@FMSMSX37>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

"Balani, Sandeep" wrote:
> 
> Hi Qiaobing,
>     I think the reason for asking such question was we have ADDIP & DELETEIP
> procedure on an Association.
> 
> Now Assume we have an Endpoint
> 
> E1 =  [ IP-A, IP-B, and Port-X ] .
> 
> Now lets say i have 2 associations on endpoint E1.
> 
> A1 = [ Endpoint-E1 , DstAddr= D1, DstPort-P1 ]
> A2 = [ Endpoint-E2, DstAddr=D2,  DstPost-P2 ]
> 
> Now when i do an  ADDIP [ for Address IP-C ] on association A1, actually i
> am adding a source  IP address in Endpoint E1 i.e. [ IP-A, IP-B, IP-C,
> Port-X ], so that I will find an Endpoint for an incoming packet.
> 
> But IP-C is only used by Association A1, Not by A2.  

You imply here that, after the ADDIP call on E1, IP-C will start to
appear as the src addr in some packets sent to D1 over A1, but will
_never_ show up in packets sent to D2 by E2 over A2. If so, there should
be no problem and everything should work as intended.

As far as the above behavior is guaranteed, you have implemented it
right.

> That's the reason we
> have an Association using Subset of IP addresses from an Endpoint.
> 
> My Question is : Is it the intended behaviour in ADDIP/DELETEIP case ????

ditto.

-Qiaobing

> 
> Many thanks
> sandeep
> -----Original Message-----
> From: Qiaobing Xie [mailto:qxie1@email.mot.com]
> Sent: Tuesday, April 23, 2002 7:50 AM
> To: Khan, Mohammad
> Cc: tsvwg@ietf.org; Pillai, Hari
> Subject: Re: [Tsvwg] End point and association Address conflict
> 
> Hi, Kashif,
> 
> Yes, you are allowed to do so, as long as you can guarantee that this
> endpoint will never use Z as the source IP address in any packet sent
> over this association. This is not hard to prevent if the endpoint has
> full control over the source IP address selection routine.
> 
> -Qiaobing
> 
> "Khan, Mohammad" wrote:
> >
> > Hi All,
> > I have following querry regarding endpoint and association source
> addresses.
> >
> > Suppose an Endpoint has three addresses X, Y and Z.
> >         An association which uses this endpoint and specifies only X and Y
> > addresses as source IP addresses. Is it allowed?
> >
> > Regards
> > kashif
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Apr 27 17:15:41 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA24636
	for <tsvwg-archive@odin.ietf.org>; Sat, 27 Apr 2002 17:15:41 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA28992
	for tsvwg-archive@odin.ietf.org; Sat, 27 Apr 2002 17:15:44 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA26999;
	Sat, 27 Apr 2002 16:48:42 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA25678
	for <tsvwg@optimus.ietf.org>; Fri, 26 Apr 2002 05:00:55 -0400 (EDT)
Received: from wiproecmx2.wipro.com (wiproecmx2.wipro.com [164.164.31.6])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA27703
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 05:00:47 -0400 (EDT)
Received: from ecvwall1.wipro.com (venus.easi.soft.net [164.164.23.6])
	by wiproecmx2.wipro.com (8.11.3/8.11.3) with SMTP id g3Q8ww502944
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 14:28:58 +0530 (IST)
Received: from wipro.com ([192.168.178.83]) by
          ecmail.mail.wipro.com (Netscape Messaging Server 4.15) with
          ESMTP id GV5ZS900.7VZ; Fri, 26 Apr 2002 13:06:09 +0530 
Message-ID: <3CC90411.9EDDB3EC@wipro.com>
Date: Fri, 26 Apr 2002 13:08:57 +0530
From: Mohammed Ashraf <mohammed.asharaf@wipro.com>
Organization: Wipro
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.4.2-2 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Dave.Sclarsky@radisys.com
CC: tsvwg@ietf.org, Simon.Zhuang@radisys.com, Heinz.Prantner@radisys.com,
        Gene.Litt@radisys.com, Seamus.Gilchrist@radisys.com
Subject: Re: [Tsvwg] T3 cannot timeout???
References: <OFE3B784B7.884E9280-ON85256BA7.0016B1AC@radisys.com>
Content-Type: multipart/mixed;
	boundary="----=_NextPartTM-000-016ccb87-58ea-11d6-a942-00b0d0d06be8"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


This is a multi-part message in MIME format.

------=_NextPartTM-000-016ccb87-58ea-11d6-a942-00b0d0d06be8
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit


Dave.Sclarsky@radisys.com wrote:

> Hello,
> We've found a Here's a case we found, hopefully someone can help!
>
> 2 peers (A and B) with 2 destinations (1 and 2) each, both using the same
> path as the 'primary'.
> Primary ethernet goes down.
> Side A detects this, and switches to destination 2.
> Side B's T3 timer hasn't yet expired, but starts receiving data from A on
> 2.  B sends SACK on 2nd path (since that's where it got the data), and
> since it also has data to send, it bundles DATA chunks with the SACK.
> The data bundled with the (many?) SACKs on dest 2 will eventually be
> SACKed.  These SACKs must contain GAP reports, indicating the missing TSNs
> (the ones sent to dest 1 with TSNs lower than those bundled with the
> SACKs).  Once 4 of these SACKs with GAPs are received, Side B will "fast
> retransmit" those chunks on dest 2.  According to the RFC (7.2.4), Side B
> will restart T3 for dest 1, since it will be retransmitting the lowest TSN
> sent to it.  It appears that this will continue forever.  Side B will never
> see dest 1 as failed, since its T3 will continually be reset, but it will
> waste a lot of time and energy sending it data!
>

Shudn't the timer be started against the destination on which this lowest
TSN is retransmitted ( i.e here Dest :2  )?



> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg


------=_NextPartTM-000-016ccb87-58ea-11d6-a942-00b0d0d06be8
Content-Type: text/plain;
	name="Wipro_Disclaimer.txt"
Content-Disposition: attachment;
	filename="Wipro_Disclaimer.txt"
Content-Transfer-Encoding: 7bit

**************************Disclaimer************************************
      


Information contained in this E-MAIL being proprietary to Wipro Limited
is 'privileged' and 'confidential' and intended for use only by the
individual or entity to which it is addressed. You are notified that any
use, copying or dissemination of the information contained in the E-MAIL
in any manner whatsoever is strictly prohibited.



 ********************************************************************

------=_NextPartTM-000-016ccb87-58ea-11d6-a942-00b0d0d06be8--


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Apr 27 17:20:53 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA25235
	for <tsvwg-archive@odin.ietf.org>; Sat, 27 Apr 2002 17:20:53 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA29061
	for tsvwg-archive@odin.ietf.org; Sat, 27 Apr 2002 17:20:57 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA27246;
	Sat, 27 Apr 2002 16:50:22 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA01691
	for <tsvwg@ns.ietf.org>; Fri, 26 Apr 2002 19:38:06 -0400 (EDT)
Received: from ftpbox.mot.com (ftpbox.mot.com [129.188.136.101])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA28805
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 19:38:03 -0400 (EDT)
Received: [from pobox.mot.com (pobox.mot.com [129.188.137.100]) by ftpbox.mot.com (ftpbox 2.1) with ESMTP id QAA00481; Fri, 26 Apr 2002 16:38:05 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox.mot.com (MOT-pobox 2.0) with ESMTP id QAA06171; Fri, 26 Apr 2002 16:38:04 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3QNc1j24993;
	Fri, 26 Apr 2002 18:38:01 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id SAA17306;
	Fri, 26 Apr 2002 18:38:09 -0500 (CDT)
Message-ID: <3CC9E579.7DB84355@email.mot.com>
Date: Fri, 26 Apr 2002 18:40:41 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: "Balani, Sandeep" <s_balani@trillium.com>
CC: "'Qiaobing Xie'" <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, randall@stewart.chicago.il.us,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F1CE15E08172D4119247009027AE9D50191173A9@FMSMSX37>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi, Sandeep,

> [ Sandeep ] I totally totally agree with U on that " Addition/Deletion of IP
> Address should be on an Endpoint & not on Association.  After all we are
> modifying the source IP addresses which is an essential part of Endpoint "
> 
> But Currently ADDIP/DELETEIP Draft Says :
> 
> "
> - Dynamic IP address reconfiguration extension: For
>     computational or networking platforms that allow addition/removal of
>     physical interface cards this feature can provide:
> 
>      A) a graceful method to add to the interfaces of an existing
>         *******association*******. For IPv6 this feature allows renumbering
>         of existing associations.
> "

I agree with you. These are loose wordings which may need to be
tightened up in the draft. But in a general sense, with the correct
understanding and clarification of the rules, the above wording is still
fine - it simply depicts the final beneficiary of this add/delete-ip
feature, which is indeed the associations themselves.

More importantly, in the draft we need to explicit notify the
implementor that using add-ip/delete-ip is fine and benefiting, but one
must not break the association rules, otherwise, it becomes a misuse of
a good tool and you implementation will break.

I am glad this is brought up in this discussion thread. 

> 
> There are similar references like this in the DRAFT, which mentions
> ADD/DELETE IP procedures assocaitions-wise.

We will improve the text and make it tighter.

> So Is there any plans to change the defination of ADDIP/DELETEIP because If
> i would do ADD/DELETE on Endpoint I have to perform ASCONF/ASCONF-ACK
> procedures on all the assocations  belonging to that endpoint which seems
> correct to me but In that case I won't be  in compliance with draft  ....

I don't think anything needs to be re-defined. We just need more
clarifications. To the implementor, a good add/delete-ip API semantics
might help to provide a safeguard against misuses. 

SCTP is different from TCP in that it allows a single endpoint to
associate with multiple peers (the "fan-out" config). But no matter how
many peers the endpoint is concurrently connected to, logically it is
_still_ a single endpoint. 

With this being cleared, it should not be hard to understand that it is
conceptually wrong to let a user apply add/delete-ip to only a subset of
the fan-out concurrent connections on that single endpoint. Because this
would in fact result in "splitting" the single endpoint into two
different ones. This is NEVER the intention of add/delete-ip - the
feature is meant to dynamically modify an endpoint without tearing down
all the associations it currently has; it is NOT meant to split a single
endpoint into two!

-Qiaobing

> 
> > With this important clarification of concept, all above should become
> > very simple and clear - what you are allowed to do here is __to delete
> > IPb from endpoint E1 which is currently being used in associations A1
> > and A9__. If your implementation is doing the right thing, the result
> > should be
> >
> >                E1'             E2
> >     A1' = [(IPa; Port1), (IPz; Port 65535)]
> >
> >     And
> >                E1'                   E2  /* call it E3 */
> >     A9' = [ (IPa; Port1), (IPm; Port 76577)]
> >
> > If your intention is to transform the original A1 and A9 to the
> > following:
> >
> >                E1'             E2
> >     A1' = [(IPa; Port1), (IPz; Port 65535)]
> >
> >     And
> >                E1'                   E2  /* call it E3 */
> >     A9' = [ (IPb; Port1), (IPm; Port 76577)]
> >
> > I am afraid that you have to tear down A1/A9 and set-up A1' and A9'
> > fresh. ADD/DELET-IP will NOT get your here without violating the basic
> > association rules.
> >
> > -Qiaobing
> >
> 
> Thanks
> Sandeep


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Apr 27 17:21:04 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA25266
	for <tsvwg-archive@odin.ietf.org>; Sat, 27 Apr 2002 17:21:03 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA29075
	for tsvwg-archive@odin.ietf.org; Sat, 27 Apr 2002 17:21:07 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA26761;
	Sat, 27 Apr 2002 16:45:33 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA04495
	for <tsvwg@ns.ietf.org>; Wed, 24 Apr 2002 18:09:49 -0400 (EDT)
Received: from motgate2.mot.com (motgate2.mot.com [136.182.1.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA10347
	for <tsvwg@ietf.org>; Wed, 24 Apr 2002 18:09:44 -0400 (EDT)
Received: [from pobox3.mot.com (pobox3.mot.com [10.64.251.242]) by motgate2.mot.com (motgate2 2.1) with ESMTP id PAA15952; Wed, 24 Apr 2002 15:08:29 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox3.mot.com (MOT-pobox3 2.0) with ESMTP id OAA15990; Wed, 24 Apr 2002 14:55:51 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3OM8Pj06189;
	Wed, 24 Apr 2002 17:08:25 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id RAA25861;
	Wed, 24 Apr 2002 17:08:32 -0500 (CDT)
Message-ID: <3CC72D6B.2BD11868@email.mot.com>
Date: Wed, 24 Apr 2002 17:10:51 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Randall Stewart <randall@stewart.chicago.il.us>
CC: Ivan.Arias-Rodriguez@nokia.com, s_balani@trillium.com,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>, mo_khan@trillium.com,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFA8@esebe017.NOE.Nokia.com> <3CC7176A.DAA9D0A0@stewart.chicago.il.us>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

...
> >         However, when you create an association, it initially uses ALL the IP addresses of that endpoint (no matter if your user started it or if the peer was the one sending the INIT chunk)... This is my understanding... :-/ If you create the association in a different endpoint, you can use a different set of addresses, though...
> 
> I agree 100%.. when you send out your INIT or INIT-ACK you MUST USE ALL
> of your addresses that are BOUND to the endpoint.. You should NEVER
> send a "sub-set" of the endpoints addresses..

Agree. This is a very important point. Also, when reading this
statement, one must keep in mind the difference between "ALL addresses
BOUND to the endpoint" and "ALL addresses available to the OS/host."
They are not always the same. And with IPv6 and its fancy scoping rules,
they probably will always be different. 

-Qiaobing


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Apr 27 17:21:30 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA25291
	for <tsvwg-archive@odin.ietf.org>; Sat, 27 Apr 2002 17:21:30 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA29105
	for tsvwg-archive@odin.ietf.org; Sat, 27 Apr 2002 17:21:33 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA26927;
	Sat, 27 Apr 2002 16:48:08 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA22432
	for <tsvwg@ns.ietf.org>; Thu, 25 Apr 2002 20:42:07 -0400 (EDT)
Received: from motgate.mot.com (motgate.mot.com [129.188.136.100])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA10153
	for <tsvwg@ietf.org>; Thu, 25 Apr 2002 20:42:03 -0400 (EDT)
Received: [from pobox4.mot.com (pobox4.mot.com [10.64.251.243]) by motgate.mot.com (motgate 2.1) with ESMTP id RAA14084; Thu, 25 Apr 2002 17:42:05 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox4.mot.com (MOT-pobox4 2.0) with ESMTP id RAA24772; Thu, 25 Apr 2002 17:42:04 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3Q0g1j23428;
	Thu, 25 Apr 2002 19:42:01 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id TAA02391;
	Thu, 25 Apr 2002 19:42:08 -0500 (CDT)
Message-ID: <3CC8A2F2.5AD7D7ED@email.mot.com>
Date: Thu, 25 Apr 2002 19:44:34 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: randall@stewart.chicago.il.us, s_balani@trillium.com,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>, mo_khan@trillium.com,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFAE@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Ivan.Arias-Rodriguez@nokia.com wrote:
....
>         Now I have another doubt... Imagine that we have an endpoint 
> E1 with addresses IPa, IPb and port 1. We have an association A1 
> that only uses IPa, whose peer is IPz and port 65535. Thus the 
> association A1 can be represented as [(IPa; Port1), (IPz; Port 65535)]. 
> What happens now if we receive an INIT from IPz/65535 directed to IPb/1?
> 
>         As IPb is not being used in A1, we should answer with an 
> INIT ACK as in a normal establishment of an association. However, 
> when you receive an INIT chunk, you include all the addresses of the endpoint
                                      ^^^^^^^^^^^^^
Here your "..include all the addr..." directly violated the "under no
circumstance can an SCTP transport address ever be present in more than
one SCTP endpoint" rule (see last the paragraph, page 23 of the SCTP
book).

        E1               E2
A1 = [(IPa; Port1), (IPz; Port 65535)]

        E3                 E2
A2 = [(IPa, IPb; Port1), (IPz; Port 65535)]

And you have IPa/Port1 present in both E1 and E3 (they are TWO different
endpoints by definition!). This is prohibited by the SCTP association
definition.

-Qiaobing

> in the INIT ACK chunk... But we have to be careful, because in
> this case, when receiving the COOKIE ECHO, we have two options:


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Apr 27 17:22:53 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA25473
	for <tsvwg-archive@odin.ietf.org>; Sat, 27 Apr 2002 17:22:53 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA29214
	for tsvwg-archive@odin.ietf.org; Sat, 27 Apr 2002 17:22:57 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA27188;
	Sat, 27 Apr 2002 16:50:06 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA27739
	for <tsvwg@ns.ietf.org>; Fri, 26 Apr 2002 18:21:43 -0400 (EDT)
Received: from motgate2.mot.com (motgate2.mot.com [136.182.1.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA20246
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 18:21:38 -0400 (EDT)
Received: [from pobox.mot.com (pobox.mot.com [129.188.137.100]) by motgate2.mot.com (motgate2 2.1) with ESMTP id PAA14879; Fri, 26 Apr 2002 15:21:38 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox.mot.com (MOT-pobox 2.0) with ESMTP id PAA29419; Fri, 26 Apr 2002 15:21:38 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3QMLUj20269;
	Fri, 26 Apr 2002 17:21:30 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id RAA16481;
	Fri, 26 Apr 2002 17:21:37 -0500 (CDT)
Message-ID: <3CC9D38A.C13724FC@email.mot.com>
Date: Fri, 26 Apr 2002 17:24:10 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: "Balani, Sandeep" <s_balani@trillium.com>
CC: "'Qiaobing Xie'" <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, randall@stewart.chicago.il.us,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F1CE15E08172D4119247009027AE9D50191173A8@FMSMSX37>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi, Sandeep,

It is very good point. Please see below for my answer.

> [ Sandeep ]  A slight change in the example above. What if initally I have 2
> Associations.
> 
>            E1                  E2
>   A1 = [(IPa,IPb; Port1), (IPz; Port 65535)]
> 
>               E1                      E2
>   A9 = [ (IPa, IPb; Port1), (IPm; Port 76577)]
>
> And then i perform DELETEIP(IPb) from association A1,  then i have something
> like
> 
>           E1'                  E2
>     A1' = [(IPa; Port1), (IPz; Port 65535)]
> 
>     And
>                  E1                      E2  /* call it E3 */
>     A9 = [ (IPa, IPb; Port1), (IPm; Port 76577)]
> 
>  Which is also illegal ....

Yes, this is illegal. Your implementation should never allow this to
happen. 

You may have some confusion over an important definition in SCTP. It is
conceptually WRONG to say "deleting IPb from association A1". One can
only add or delete addresses to or from __an SCTP endpoint__. 

With this important clarification of concept, all above should become
very simple and clear - what you are allowed to do here is __to delete
IPb from endpoint E1 which is currently being used in associations A1
and A9__. If your implementation is doing the right thing, the result
should be 

               E1'             E2
    A1' = [(IPa; Port1), (IPz; Port 65535)]

    And
               E1'                   E2  /* call it E3 */
    A9' = [ (IPa; Port1), (IPm; Port 76577)]

If your intention is to transform the original A1 and A9 to the
following:

               E1'             E2
    A1' = [(IPa; Port1), (IPz; Port 65535)]

    And
               E1'                   E2  /* call it E3 */
    A9' = [ (IPb; Port1), (IPm; Port 76577)]

I am afraid that you have to tear down A1/A9 and set-up A1' and A9'
fresh. ADD/DELET-IP will NOT get your here without violating the basic
association rules.

-Qiaobing


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Apr 27 17:30:20 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA26462
	for <tsvwg-archive@odin.ietf.org>; Sat, 27 Apr 2002 17:30:19 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA29384
	for tsvwg-archive@odin.ietf.org; Sat, 27 Apr 2002 17:30:23 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA27069;
	Sat, 27 Apr 2002 16:49:27 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA19634
	for <tsvwg@ns.ietf.org>; Fri, 26 Apr 2002 15:51:05 -0400 (EDT)
Received: from motgate4.mot.com (motgate4.mot.com [144.189.100.102])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA02405
	for <tsvwg@ietf.org>; Fri, 26 Apr 2002 15:51:01 -0400 (EDT)
Received: [from pobox.mot.com (pobox.mot.com [129.188.137.100]) by motgate4.mot.com (motgate4 2.1) with ESMTP id MAA19869 for <tsvwg@ietf.org>; Fri, 26 Apr 2002 12:51:04 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox.mot.com (MOT-pobox 2.0) with ESMTP id MAA09630 for <tsvwg@ietf.org>; Fri, 26 Apr 2002 12:51:03 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3QJouj08216;
	Fri, 26 Apr 2002 14:50:56 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id OAA14852;
	Fri, 26 Apr 2002 14:51:02 -0500 (CDT)
Message-ID: <3CC9B03F.444E9EAF@email.mot.com>
Date: Fri, 26 Apr 2002 14:53:35 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: Qiaobing Xie <Qiaobing.Xie@motorola.com>, randall@stewart.chicago.il.us,
        s_balani@trillium.com, mo_khan@trillium.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFAF@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Hi, Ivan,

This is a very important discussion. Please see my comments below...

Ivan.Arias-Rodriguez@nokia.com wrote:
....
> > Here your "..include all the addr..." directly violated the "under no
> > circumstance can an SCTP transport address ever be present in
> > more than
> > one SCTP endpoint" rule (see last the paragraph, page 23 of the SCTP
> > book).
> >
> >         E1               E2
> > A1 = [(IPa; Port1), (IPz; Port 65535)]
> >
> >         E3                 E2
> > A2 = [(IPa, IPb; Port1), (IPz; Port 65535)]
> >
> > And you have IPa/Port1 present in both E1 and E3 (they are TWO different 
> > endpoints by definition!). This is prohibited by the SCTP association
> > definition.
> 
>         No, I don't think what I'm proposing is illegal at all... Take into account that you can make use of the AddIP extension to delete IPb... Thus, initially, the association had all the addresses of the endpoint (IPa and IPb). Later on you delete your source address IPb. And then all this happens...

Applying Add/Delete-IP is ok, but anytime you do that, you must not
break the basic rules that define the endpoint and association.
Otherwise, you will get into troubles as you've noticed in your own
analysis.

Let us layout what you are proposing in details.

Suppose you originally have:

          E1                  E2
  A1 = [(IPa,IPb; Port1), (IPz; Port 65535)]

then you delete IPb from E1 (Note, this effectively changes the
definition of E1 to E1' and A1 to A1'. They are logically differently
defined now):

          E1'                  E2
  A1' = [(IPa; Port1), (IPz; Port 65535)]

Now, E2 sends an INIT to IPb/Port1, and you decide it's ok to setup
assoc A2, which is ok. But you are _PROBINITED_ by the endpoint
definition to have:

          E1'                  E2
  A1' = [(IPa; Port1), (IPz; Port 65535)]

          E3                  E2
  A2 = [(IPa,IPb; Port1), (IPz; Port 65535)]  /* bad, illegal assoc */

I don't know how you can detect and prevent this illegal setup conflict
in your implementation. But allowing this will violate the basic SCTP
association rule I mentioned above. Another way to understand why this
is illegal is to think about this: if this were to be allowed, what
would happen if you then go ahead and add IPb back to A1'? You would
magically end up with _two_ association between the same pair of
endpoints!?! 

However, it is perfectly ok to have:

          E1'                  E2
  A1' = [(IPa; Port1), (IPz; Port 65535)]

          E3                  E2
  A2 = [(IPb; Port1), (IPz; Port 65535)]  /* good, legal assoc */

This discussion is very good. We should capture this and caution people
not to break SCTP association definitions when using ADD/DELETE-IP.

-Qiaobing

> 
>         So, the problem is that the pair IPb/1-IPz/65535 is not part of any association... :-/
> 
>         BR Iván Arias Rodríguez
> 
> > -Qiaobing
> >
> > > in the INIT ACK chunk... But we have to be careful, because in
> > > this case, when receiving the COOKIE ECHO, we have two options:
> >


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 02:46:59 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA03500
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 02:46:59 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id CAA28554
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 02:47:00 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id CAA27128;
	Mon, 29 Apr 2002 02:06:11 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id CAA27099
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 02:06:09 -0400 (EDT)
Received: from caduceus.fm.intel.com (fmr02.intel.com [192.55.52.25])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA25757
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 02:06:05 -0400 (EDT)
Received: from talaria.fm.intel.com (talaria.fm.intel.com [10.1.192.39])
	by caduceus.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.42 2002/04/26 23:25:13 root Exp $) with ESMTP id g3T65Cn01184
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 06:05:12 GMT
Received: from fmsmsxvs041.fm.intel.com (fmsmsxv041-1.fm.intel.com [132.233.48.109])
	by talaria.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.17 2002/04/27 00:24:04 root Exp $) with SMTP id g3T68NM18074
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 06:08:23 GMT
Received: from FMSMSX017.fm.intel.com ([132.233.42.196])
 by fmsmsxvs041.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002042823062500127
 for <tsvwg@ietf.org>; Sun, 28 Apr 2002 23:06:25 -0700
Received: by fmsmsx017.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <28FQ3MWJ>; Sun, 28 Apr 2002 23:06:05 -0700
Message-ID: <A1171C2850D8D4119D7B0002A508C764016983C9@bgsmsx91.iind.intel.com>
From: "Khan, Mohammad" <mo_khan@trillium.com>
To: "'tsvwg@ietf.org'" <tsvwg@ietf.org>
Cc: "Pillai, Hari" <h_pillai@trillium.com>
Date: Sun, 28 Apr 2002 23:03:50 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: [Tsvwg] EndPoint and Association Address conflict
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi All,
I could make the following observations from the ongoing discussions about
Endpoint and Association address conflict for SCTP.

1)INIT chunk shall contain all the addresses bound to the Endpoint.

2)Addition and deletion of IP address shall be on Endpoint not on
Association.

For the point 2) suppose an Endpoint E1(IPA and Port X) has multiple
associations A1, A2 and A3 with three different peers.

	E1(IPA and PortX)			Peer Hosts
	A1(E1)					IPB and Port X
	A2(E1)					IPC and Port X
	A3(E1)					IPD and Port X
As per the assumption 2) to add an IPZ at E1 SCTP has to send ASCONF chunk
on all the A1, A2 and A3 associations and it shall not consider this address
as added till it receives success from all the associations. If failure is
received is received from any of the associations then this Add-IP request
shall be discarded. 

Same procedure is to be adopted to delete an address from an endpoint.
Please comment.

Regards
kashif

		

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 03:20:26 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA06106
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 03:20:26 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id DAA00602
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 03:20:30 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id CAA28671;
	Mon, 29 Apr 2002 02:50:13 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id CAA28642
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 02:50:10 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-230.dallas.net [209.44.42.230])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA03728
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 02:50:04 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id BAA17877;
	Mon, 29 Apr 2002 01:49:31 -0500
Date: Mon, 29 Apr 2002 01:49:31 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Qiaobing Xie <qxie1@email.mot.com>
Cc: "Balani, Sandeep" <s_balani@trillium.com>,
        "'Qiaobing Xie'" <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, randall@stewart.chicago.il.us,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
Message-ID: <20020429014931.A17185@openss7.org>
Reply-To: bidulock@openss7.org
References: <F1CE15E08172D4119247009027AE9D50191173A9@FMSMSX37> <3CC9E579.7DB84355@email.mot.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <3CC9E579.7DB84355@email.mot.com>; from qxie1@email.mot.com on Fri, Apr 26, 2002 at 06:40:41PM -0500
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Qiaobing,

Multi-homing is OPTIONAL.  Are you saying that a group of
associations that are not multi-homed (single IP address
in INIT), and on different IP addresses for the same host
is somehow splitting the host into multiple end-points?
I don't see how this is different than for TCP when the
host has multiple IP addresses: each connection may be
from a different IP address...  Please explain.

--brian

On Fri, 26 Apr 2002, Qiaobing Xie wrote:

> 
> I don't think anything needs to be re-defined. We just need more
> clarifications. To the implementor, a good add/delete-ip API semantics
> might help to provide a safeguard against misuses. 
> 
> SCTP is different from TCP in that it allows a single endpoint to
> associate with multiple peers (the "fan-out" config). But no matter how
> many peers the endpoint is concurrently connected to, logically it is
> _still_ a single endpoint. 
> 
> With this being cleared, it should not be hard to understand that it is
> conceptually wrong to let a user apply add/delete-ip to only a subset of
> the fan-out concurrent connections on that single endpoint. Because this
> would in fact result in "splitting" the single endpoint into two
> different ones. This is NEVER the intention of add/delete-ip - the
> feature is meant to dynamically modify an endpoint without tearing down
> all the associations it currently has; it is NOT meant to split a single
> endpoint into two!
> 
> -Qiaobing
> 

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 03:43:54 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA08098
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 03:43:54 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id DAA01924
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 03:43:58 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id CAA28841;
	Mon, 29 Apr 2002 02:56:38 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id CAA28811
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 02:56:35 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-230.dallas.net [209.44.42.230])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA04280
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 02:56:27 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id BAA17983;
	Mon, 29 Apr 2002 01:56:20 -0500
Date: Mon, 29 Apr 2002 01:56:20 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: "Khan, Mohammad" <mo_khan@trillium.com>
Cc: "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
Message-ID: <20020429015620.B17185@openss7.org>
Reply-To: bidulock@openss7.org
References: <A1171C2850D8D4119D7B0002A508C764016983C9@bgsmsx91.iind.intel.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <A1171C2850D8D4119D7B0002A508C764016983C9@bgsmsx91.iind.intel.com>; from mo_khan@trillium.com on Sun, Apr 28, 2002 at 11:03:50PM -0700
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Khan,,

On Sun, 28 Apr 2002, Khan, Mohammad wrote:

> Hi All,
> I could make the following observations from the ongoing discussions about
> Endpoint and Association address conflict for SCTP.
> 
> 1)INIT chunk shall contain all the addresses bound to the Endpoint.

RFC 2960 permits non-multihoming.  It also permits the addresses in
the INIT to be a subset of the addresses available to an Endpoint.
A user process starting an association might not have permission to
all the addresses owned by a host.  How do you propose to make this
requirement stick?

If you mean that when an association happens to be bound with
INADDR_ANY and an IP is added, and the host determines in some
implementation dependent fashion that a new dynamic address should
be subsequently added to this association, that it should perform
the ADDIP procedures on behalf of the association, then there might
not be a problem.  But the above seems to rule out many features of
SCTP that are already implemented and being used.

> 
> 2)Addition and deletion of IP address shall be on Endpoint not on
> Association.
> 
> For the point 2) suppose an Endpoint E1(IPA and Port X) has multiple
> associations A1, A2 and A3 with three different peers.
> 
> 	E1(IPA and PortX)			Peer Hosts
> 	A1(E1)					IPB and Port X
> 	A2(E1)					IPC and Port X
> 	A3(E1)					IPD and Port X
> As per the assumption 2) to add an IPZ at E1 SCTP has to send ASCONF chunk
> on all the A1, A2 and A3 associations and it shall not consider this address
> as added till it receives success from all the associations. If failure is
> received is received from any of the associations then this Add-IP request
> shall be discarded. 

Huh?  Why must it fail for all if it fails for one?  That doesn't
sound too wise.  For example, if the receiver of the ASCONF does not
like the new destination (perhaps it does not meet security criteria)
it would not be able to reject it?

--brian

> 
> Same procedure is to be adopted to delete an address from an endpoint.
> Please comment.
> 
> Regards
> kashif
> 
> 		
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 04:04:52 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA10040
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 04:04:51 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id EAA03203
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 04:04:53 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA01517;
	Mon, 29 Apr 2002 03:32:23 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA01488
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 03:32:20 -0400 (EDT)
Received: from hermes.fm.intel.com (fmr01.intel.com [192.55.52.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA07019
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 03:32:15 -0400 (EDT)
Received: from talaria.fm.intel.com (talaria.fm.intel.com [10.1.192.39])
	by hermes.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.42 2002/04/26 23:25:13 root Exp $) with ESMTP id g3T7XEh14231
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 07:33:14 GMT
Received: from fmsmsxvs042.fm.intel.com (fmsmsxv042-1.fm.intel.com [132.233.48.110])
	by talaria.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.17 2002/04/27 00:24:04 root Exp $) with SMTP id g3T7YXw17168
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 07:34:33 GMT
Received: from fmsmsx26.fm.intel.com ([132.233.42.26])
 by fmsmsxvs042.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002042900363309778
 ; Mon, 29 Apr 2002 00:36:33 -0700
Received: by fmsmsx26.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <2HL0RRZ5>; Mon, 29 Apr 2002 00:32:15 -0700
Message-ID: <A1171C2850D8D4119D7B0002A508C764016983CA@bgsmsx91.iind.intel.com>
From: "Khan, Mohammad" <mo_khan@trillium.com>
To: "'bidulock@openss7.org'" <bidulock@openss7.org>,
        "Khan, Mohammad" <mo_khan@trillium.com>
Cc: "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: RE: [Tsvwg] EndPoint and Association Address conflict
Date: Mon, 29 Apr 2002 00:30:00 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id DAA01489
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Hi Brian,

-----Original Message-----
From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
Sent: Monday, April 29, 2002 12:26 PM
To: Khan, Mohammad
Cc: 'tsvwg@ietf.org'; Pillai, Hari
Subject: Re: [Tsvwg] EndPoint and Association Address conflict


Khan,,

On Sun, 28 Apr 2002, Khan, Mohammad wrote:

> Hi All,
> I could make the following observations from the ongoing discussions about
> Endpoint and Association address conflict for SCTP.
> 
> 1)INIT chunk shall contain all the addresses bound to the Endpoint.

RFC 2960 permits non-multihoming.  It also permits the addresses in
the INIT to be a subset of the addresses available to an Endpoint.
A user process starting an association might not have permission to
all the addresses owned by a host.  How do you propose to make this
requirement stick?

If you mean that when an association happens to be bound with
INADDR_ANY and an IP is added, and the host determines in some
implementation dependent fashion that a new dynamic address should
be subsequently added to this association, that it should perform
the ADDIP procedures on behalf of the association, then there might
not be a problem.  But the above seems to rule out many features of
SCTP that are already implemented and being used.

This is what Randell mentioned in one of his comments
"
Why did you not include IPb in the INIT or INIT-ACK that setup A1. You
are
supposed to tell the peer of all your addresses that are bound to your
endpoint. 
You really MUST tell the peer of all your endpoint's addresses.  
R
"
> 
> 2)Addition and deletion of IP address shall be on Endpoint not on
> Association.
> 
> For the point 2) suppose an Endpoint E1(IPA and Port X) has multiple
> associations A1, A2 and A3 with three different peers.
> 
> 	E1(IPA and PortX)			Peer Hosts
> 	A1(E1)					IPB and Port X
> 	A2(E1)					IPC and Port X
> 	A3(E1)					IPD and Port X
> As per the assumption 2) to add an IPZ at E1 SCTP has to send ASCONF chunk
> on all the A1, A2 and A3 associations and it shall not consider this
address
> as added till it receives success from all the associations. If failure is
> received is received from any of the associations then this Add-IP request
> shall be discarded. 

Huh?  Why must it fail for all if it fails for one?  That doesn't
sound too wise.  For example, if the receiver of the ASCONF does not
like the new destination (perhaps it does not meet security criteria)
it would not be able to reject it? 

Seems quite annoyed. If u are accepting ADD/DELETE IP as successfull even if
ADD/DELETE IP is not successful for all the associations then again u will
end up in having two instances of one transport address between two
associations of same endpoint. That is what we all had been discussing till
now.
---kashif

--brian


> 
> Same procedure is to be adopted to delete an address from an endpoint.
> Please comment.
> 
> Regards
> kashif
> 
> 		
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 04:26:45 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA11576
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 04:26:45 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id EAA04211
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 04:26:47 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA02502;
	Mon, 29 Apr 2002 03:58:26 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA02476
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 03:58:24 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-230.dallas.net [209.44.42.230])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA09445
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 03:58:19 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id CAA18724;
	Mon, 29 Apr 2002 02:57:54 -0500
Date: Mon, 29 Apr 2002 02:57:54 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: "Khan, Mohammad" <mo_khan@trillium.com>
Cc: "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
Message-ID: <20020429025754.A15434@openss7.org>
Reply-To: bidulock@openss7.org
References: <A1171C2850D8D4119D7B0002A508C764016983CA@bgsmsx91.iind.intel.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <A1171C2850D8D4119D7B0002A508C764016983CA@bgsmsx91.iind.intel.com>; from mo_khan@trillium.com on Mon, Apr 29, 2002 at 12:30:00AM -0700
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Mohammad,

Please see comments below...

On Mon, 29 Apr 2002, Khan, Mohammad wrote:

> Hi Brian,
> 
> -----Original Message-----
> From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> Sent: Monday, April 29, 2002 12:26 PM
> To: Khan, Mohammad
> Cc: 'tsvwg@ietf.org'; Pillai, Hari
> Subject: Re: [Tsvwg] EndPoint and Association Address conflict
> 
> 
> Khan,,
> 
> On Sun, 28 Apr 2002, Khan, Mohammad wrote:
> 
> > Hi All,
> > I could make the following observations from the ongoing discussions about
> > Endpoint and Association address conflict for SCTP.
> > 
> > 1)INIT chunk shall contain all the addresses bound to the Endpoint.
> 
> RFC 2960 permits non-multihoming.  It also permits the addresses in
> the INIT to be a subset of the addresses available to an Endpoint.
> A user process starting an association might not have permission to
> all the addresses owned by a host.  How do you propose to make this
> requirement stick?
> 
> If you mean that when an association happens to be bound with
> INADDR_ANY and an IP is added, and the host determines in some
> implementation dependent fashion that a new dynamic address should
> be subsequently added to this association, that it should perform
> the ADDIP procedures on behalf of the association, then there might
> not be a problem.  But the above seems to rule out many features of
> SCTP that are already implemented and being used.
> 
> This is what Randell mentioned in one of his comments
> "
> Why did you not include IPb in the INIT or INIT-ACK that setup A1. You
> are supposed to tell the peer of all your addresses that are bound to
> your endpoint.  You really MUST tell the peer of all your endpoint's
> addresses.  
> R
> "

Doesn't say that in RFC 2960 or anywhere else.  Here is a reason for
not including all the addresses:

I have a Signalling Gateway and an MGC, both of which have 3 interfaces:
two are high-speed interfaces on an Engineered network for signalling
applications; one is a low-speed interface which also acts as a default
route which is used for adminstrative tasks only (such as configuration,
SNMP, etc).  Although the SG can talk to the MGC on the default
interface, associations which are formed for signalling between the SG
and MGC can only use the interfaces dedicated to signalling on the
engineered network.

Now, it may be that the adminstrative interface is using DHCP on both
boxes.  When a DHCP lease time is renewed and a new IP address is
generated (e.g., on the SG), there is NO WAY that I am adding the new
address to the associations which are used for signalling.

An IPV6 example:  I have three interfaces which are connected to three
different service providers.  I place high-volume traffic on the two
service provided with which I can obtain the best rate for bulk traffic.
I place low-volume, low-latency traffic on the third service provider.
All 3 interfaces are DHCP.  I am NOT adding the low-volume service
provider interface to associations sending high-volume traffic.

I'm sorry but the INADDR_ANY approach does not work in many real-world
situations.


> > 
> > 2)Addition and deletion of IP address shall be on Endpoint not on
> > Association.
> > 
> > For the point 2) suppose an Endpoint E1(IPA and Port X) has multiple
> > associations A1, A2 and A3 with three different peers.
> > 
> > 	E1(IPA and PortX)			Peer Hosts
> > 	A1(E1)					IPB and Port X
> > 	A2(E1)					IPC and Port X
> > 	A3(E1)					IPD and Port X
> > As per the assumption 2) to add an IPZ at E1 SCTP has to send ASCONF chunk
> > on all the A1, A2 and A3 associations and it shall not consider this
> address
> > as added till it receives success from all the associations. If failure is
> > received is received from any of the associations then this Add-IP request
> > shall be discarded. 
> 
> Huh?  Why must it fail for all if it fails for one?  That doesn't
> sound too wise.  For example, if the receiver of the ASCONF does not
> like the new destination (perhaps it does not meet security criteria)
> it would not be able to reject it? 
> 
> Seems quite annoyed. If u are accepting ADD/DELETE IP as successfull even if
> ADD/DELETE IP is not successful for all the associations then again u will
> end up in having two instances of one transport address between two
> associations of same endpoint.

So what's the matter with that?  This is permitted by RFC 2960 in its
current form, with or without ADDIP.

--brian

> That is what we all had been discussing till now.
> ---kashif
> 
> --brian
> 
> 
> > 
> > Same procedure is to be adopted to delete an address from an endpoint.
> > Please comment.
> > 
> > Regards
> > kashif
> > 
> > 		
> > 
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> -- 
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 05:07:03 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA14398
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 05:07:03 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA06704
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 05:07:05 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA04045;
	Mon, 29 Apr 2002 04:22:32 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA04017
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 04:22:30 -0400 (EDT)
Received: from auemail2.firewall.lucent.com (auemail2.lucent.com [192.11.223.163])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA11269
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 04:22:27 -0400 (EDT)
Received: from horh1.emsr.lucent.com (h135-17-1-40.lucent.com [135.17.1.40])
	by auemail2.firewall.lucent.com (Switch-2.1.3/Switch-2.1.0) with ESMTP id g3T8Lqc29748;
	Mon, 29 Apr 2002 04:21:53 -0400 (EDT)
Received: from new-wopr.eng.ascend.com by horh1.emsr.lucent.com (8.9.3+Sun/EMS-1.5 Solaris/emsr)
	id EAA10474 for ; Mon, 29 Apr 2002 04:21:51 -0400 (EDT)
Received: from grigri.eng.ascend.com (grigri.eng.ascend.com [135.140.53.45])
	by new-wopr.eng.ascend.com (8.10.2+Sun/8.10.2) with ESMTP id g3T8Lpe25242;
	Mon, 29 Apr 2002 01:21:51 -0700 (PDT)
Received: from lucent.com by grigri.eng.ascend.com (8.8.8+Sun/SMI-SVR4)
	id BAA13463; Mon, 29 Apr 2002 01:21:50 -0700 (PDT)
Message-ID: <3CCD0323.46C301CF@lucent.com>
Date: Mon, 29 Apr 2002 01:24:03 -0700
From: Jacob Heitz <jheitz@lucent.com>
Organization: Lucent Technologies, Internetworking Systems, Software Engineering, 
 Alameda
X-Mailer: Mozilla 4.77 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: bidulock@openss7.org
CC: "Khan, Mohammad" <mo_khan@trillium.com>,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <A1171C2850D8D4119D7B0002A508C764016983CA@bgsmsx91.iind.intel.com> <20020429025754.A15434@openss7.org>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Brian,

You are assuming that a host is equivalent to an SCTP endpoint.
That is, that if a host has 3 IPs, then any SCTP endpoint in that
host must have the same 3 IPs. IMO, an SCTP endpoint may be
constructed with only a subset of the IPs of it's host. ie,
this host with 3 IPs may contain an SCTP endpoint that has just
two of those IPs.

"Brian F. G. Bidulock" wrote:
> 
> Mohammad,
> 
> Please see comments below...
> 
> On Mon, 29 Apr 2002, Khan, Mohammad wrote:
> 
> > Hi Brian,
> >
> > -----Original Message-----
> > From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> > Sent: Monday, April 29, 2002 12:26 PM
> > To: Khan, Mohammad
> > Cc: 'tsvwg@ietf.org'; Pillai, Hari
> > Subject: Re: [Tsvwg] EndPoint and Association Address conflict
> >
> >
> > Khan,,
> >
> > On Sun, 28 Apr 2002, Khan, Mohammad wrote:
> >
> > > Hi All,
> > > I could make the following observations from the ongoing discussions about
> > > Endpoint and Association address conflict for SCTP.
> > >
> > > 1)INIT chunk shall contain all the addresses bound to the Endpoint.
> >
> > RFC 2960 permits non-multihoming.  It also permits the addresses in
> > the INIT to be a subset of the addresses available to an Endpoint.
> > A user process starting an association might not have permission to
> > all the addresses owned by a host.  How do you propose to make this
> > requirement stick?
> >
> > If you mean that when an association happens to be bound with
> > INADDR_ANY and an IP is added, and the host determines in some
> > implementation dependent fashion that a new dynamic address should
> > be subsequently added to this association, that it should perform
> > the ADDIP procedures on behalf of the association, then there might
> > not be a problem.  But the above seems to rule out many features of
> > SCTP that are already implemented and being used.
> >
> > This is what Randell mentioned in one of his comments
> > "
> > Why did you not include IPb in the INIT or INIT-ACK that setup A1. You
> > are supposed to tell the peer of all your addresses that are bound to
> > your endpoint.  You really MUST tell the peer of all your endpoint's
> > addresses.
> > R
> > "
> 
> Doesn't say that in RFC 2960 or anywhere else.  Here is a reason for
> not including all the addresses:
> 
> I have a Signalling Gateway and an MGC, both of which have 3 interfaces:
> two are high-speed interfaces on an Engineered network for signalling
> applications; one is a low-speed interface which also acts as a default
> route which is used for adminstrative tasks only (such as configuration,
> SNMP, etc).  Although the SG can talk to the MGC on the default
> interface, associations which are formed for signalling between the SG
> and MGC can only use the interfaces dedicated to signalling on the
> engineered network.
> 
> Now, it may be that the adminstrative interface is using DHCP on both
> boxes.  When a DHCP lease time is renewed and a new IP address is
> generated (e.g., on the SG), there is NO WAY that I am adding the new
> address to the associations which are used for signalling.
> 
> An IPV6 example:  I have three interfaces which are connected to three
> different service providers.  I place high-volume traffic on the two
> service provided with which I can obtain the best rate for bulk traffic.
> I place low-volume, low-latency traffic on the third service provider.
> All 3 interfaces are DHCP.  I am NOT adding the low-volume service
> provider interface to associations sending high-volume traffic.
> 
> I'm sorry but the INADDR_ANY approach does not work in many real-world
> situations.
> 
> > >
> > > 2)Addition and deletion of IP address shall be on Endpoint not on
> > > Association.
> > >
> > > For the point 2) suppose an Endpoint E1(IPA and Port X) has multiple
> > > associations A1, A2 and A3 with three different peers.
> > >
> > >     E1(IPA and PortX)                       Peer Hosts
> > >     A1(E1)                                  IPB and Port X
> > >     A2(E1)                                  IPC and Port X
> > >     A3(E1)                                  IPD and Port X
> > > As per the assumption 2) to add an IPZ at E1 SCTP has to send ASCONF chunk
> > > on all the A1, A2 and A3 associations and it shall not consider this
> > address
> > > as added till it receives success from all the associations. If failure is
> > > received is received from any of the associations then this Add-IP request
> > > shall be discarded.
> >
> > Huh?  Why must it fail for all if it fails for one?  That doesn't
> > sound too wise.  For example, if the receiver of the ASCONF does not
> > like the new destination (perhaps it does not meet security criteria)
> > it would not be able to reject it?
> >
> > Seems quite annoyed. If u are accepting ADD/DELETE IP as successfull even if
> > ADD/DELETE IP is not successful for all the associations then again u will
> > end up in having two instances of one transport address between two
> > associations of same endpoint.
> 
> So what's the matter with that?  This is permitted by RFC 2960 in its
> current form, with or without ADDIP.
> 
> --brian
> 
> > That is what we all had been discussing till now.
> > ---kashif
> >
> > --brian
> >
> >
> > >
> > > Same procedure is to be adopted to delete an address from an endpoint.
> > > Please comment.
> > >
> > > Regards
> > > kashif

-- 
// Jacob Heitz. Tel:510-747-2917

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 05:19:15 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA15147
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 05:19:15 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA07261
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 05:19:17 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA04517;
	Mon, 29 Apr 2002 04:30:33 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA04440
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 04:30:25 -0400 (EDT)
Received: from gorilla.mchh.siemens.de (gorilla.mchh.siemens.de [194.138.158.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA11792
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 04:30:22 -0400 (EDT)
Received: from blues.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.227])
	by gorilla.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id KAA16802;
	Mon, 29 Apr 2002 10:30:17 +0200 (MET DST)
Received: from mchh273e.demchh201e.icn.siemens.de ([139.21.200.83])
	by blues.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id KAA13935;
	Mon, 29 Apr 2002 10:29:41 +0200 (MET DST)
Received: by MCHH273E with Internet Mail Service (5.5.2653.19)
	id <JV3Q0PRT>; Mon, 29 Apr 2002 10:30:16 +0200
Message-ID: <F92978223B01D311ACFF0008C71EE19D015DF7D1@MCHH225E>
From: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
To: "'bidulock@openss7.org'" <bidulock@openss7.org>,
        Qiaobing Xie
	 <qxie1@email.mot.com>
Cc: "Balani, Sandeep" <s_balani@trillium.com>,
        "'Qiaobing Xie'"
	 <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, randall@stewart.chicago.il.us,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Mon, 29 Apr 2002 10:30:11 +0200
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id EAA04475
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Hi Brian,

if a host A has two IP-addresses IP' and IP'' and two associations
from (IP', Port) to a peer (IP, Port) and (IP'', Port) to the peer
(IP, Port) then 
- host A has two IP-addresses
- there are two end-points (IP', Port), (IP'', Port), each of them
  uses only one IP address available at host A.
This 'splits' (in your terminology) host A into two end points.
The difference between IP addresses of the end point and IP addresses
of the host is also the reason why it is valid to say
- you MUST include all IP address of the end point in the INIT INIT-ACK.
- you include a subset of all addresses of the host in the INIT INIT-ACK.

Best regards
Michael

Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> Sent:	Monday, April 29, 2002 8:50 AM
> To:	Qiaobing Xie
> Cc:	Balani, Sandeep; 'Qiaobing Xie'; Ivan.Arias-Rodriguez@nokia.com; randall@stewart.chicago.il.us; Khan, Mohammad; tsvwg@ietf.org
> Subject:	Re: [Tsvwg] End point and association Address conflict
> 
> Qiaobing,
> 
> Multi-homing is OPTIONAL.  Are you saying that a group of
> associations that are not multi-homed (single IP address
> in INIT), and on different IP addresses for the same host
> is somehow splitting the host into multiple end-points?
> I don't see how this is different than for TCP when the
> host has multiple IP addresses: each connection may be
> from a different IP address...  Please explain.
> 
> --brian
> 
> On Fri, 26 Apr 2002, Qiaobing Xie wrote:
> 
> > 
> > I don't think anything needs to be re-defined. We just need more
> > clarifications. To the implementor, a good add/delete-ip API semantics
> > might help to provide a safeguard against misuses. 
> > 
> > SCTP is different from TCP in that it allows a single endpoint to
> > associate with multiple peers (the "fan-out" config). But no matter how
> > many peers the endpoint is concurrently connected to, logically it is
> > _still_ a single endpoint. 
> > 
> > With this being cleared, it should not be hard to understand that it is
> > conceptually wrong to let a user apply add/delete-ip to only a subset of
> > the fan-out concurrent connections on that single endpoint. Because this
> > would in fact result in "splitting" the single endpoint into two
> > different ones. This is NEVER the intention of add/delete-ip - the
> > feature is meant to dynamically modify an endpoint without tearing down
> > all the associations it currently has; it is NOT meant to split a single
> > endpoint into two!
> > 
> > -Qiaobing
> > 
> 
> -- 
> Brian F. G. Bidulock    > ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 05:19:36 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA15174
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 05:19:35 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA07288
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 05:19:38 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA05448;
	Mon, 29 Apr 2002 04:44:37 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA05361
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 04:44:30 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-230.dallas.net [209.44.42.230])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA12996
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 04:44:25 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id DAA19331;
	Mon, 29 Apr 2002 03:44:06 -0500
Date: Mon, 29 Apr 2002 03:44:06 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Jacob Heitz <jheitz@lucent.com>
Cc: "Khan, Mohammad" <mo_khan@trillium.com>,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
Message-ID: <20020429034406.A19082@openss7.org>
Reply-To: bidulock@openss7.org
References: <A1171C2850D8D4119D7B0002A508C764016983CA@bgsmsx91.iind.intel.com> <20020429025754.A15434@openss7.org> <3CCD0323.46C301CF@lucent.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <3CCD0323.46C301CF@lucent.com>; from jheitz@lucent.com on Mon, Apr 29, 2002 at 01:24:03AM -0700
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Jacob,

True, in the examples I gave, the Host was one SCTP endpoint, but the
same is true for the Peer Hosts in Mohammad's example:

> > > >     E1(IPA and PortX)                       Peer Hosts
> > > >     A1(E1)                                  IPB and Port X
> > > >     A2(E1)                                  IPC and Port X
> > > >     A3(E1)                                  IPD and Port X

If one defines an Endpoint as a vector of IP addresses and a single
SCTP port number, what is to stop IP address vectors for distinct
associations from being disjoint or overlapping in any way the
implementation sees fit.  How one distinguishes endpoints is rather
arbitrary.

I don't see what is stopping this:

Before:
     Host                                    Peer Hosts
     A1(IPA and PortX)                       IPB and Port Y
     A2(IPA and PortX)                       IPC and Port Y
     A3(IPA and PortX)                       IPD and Port Y

After adding IPX, IPY, IPZ:
     Host                                    Peer Hosts
     A1(IPA, IPX and PortX)                  IPB and Port Y
     A2(IPA, IPY and PortX)                  IPC and Port Y
     A3(IPA, IPZ and PortX)                  IPD and Port Y

I could see a problem binding that way in the first place, but
for a fully formed association what does it matter?  Why define
the endpoint for the associations as EP1 just because they start
out life with the same IPA and PortX?

--brian


On Mon, 29 Apr 2002, Jacob Heitz wrote:

> Brian,
> 
> You are assuming that a host is equivalent to an SCTP endpoint.
> That is, that if a host has 3 IPs, then any SCTP endpoint in that
> host must have the same 3 IPs. IMO, an SCTP endpoint may be
> constructed with only a subset of the IPs of it's host. ie,
> this host with 3 IPs may contain an SCTP endpoint that has just
> two of those IPs.
> 
> "Brian F. G. Bidulock" wrote:
> > 
> > Mohammad,
> > 
> > Please see comments below...
> > 
> > On Mon, 29 Apr 2002, Khan, Mohammad wrote:
> > 
> > > Hi Brian,
> > >
> > > -----Original Message-----
> > > From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> > > Sent: Monday, April 29, 2002 12:26 PM
> > > To: Khan, Mohammad
> > > Cc: 'tsvwg@ietf.org'; Pillai, Hari
> > > Subject: Re: [Tsvwg] EndPoint and Association Address conflict
> > >
> > >
> > > Khan,,
> > >
> > > On Sun, 28 Apr 2002, Khan, Mohammad wrote:
> > >
> > > > Hi All,
> > > > I could make the following observations from the ongoing discussions about
> > > > Endpoint and Association address conflict for SCTP.
> > > >
> > > > 1)INIT chunk shall contain all the addresses bound to the Endpoint.
> > >
> > > RFC 2960 permits non-multihoming.  It also permits the addresses in
> > > the INIT to be a subset of the addresses available to an Endpoint.
> > > A user process starting an association might not have permission to
> > > all the addresses owned by a host.  How do you propose to make this
> > > requirement stick?
> > >
> > > If you mean that when an association happens to be bound with
> > > INADDR_ANY and an IP is added, and the host determines in some
> > > implementation dependent fashion that a new dynamic address should
> > > be subsequently added to this association, that it should perform
> > > the ADDIP procedures on behalf of the association, then there might
> > > not be a problem.  But the above seems to rule out many features of
> > > SCTP that are already implemented and being used.
> > >
> > > This is what Randell mentioned in one of his comments
> > > "
> > > Why did you not include IPb in the INIT or INIT-ACK that setup A1. You
> > > are supposed to tell the peer of all your addresses that are bound to
> > > your endpoint.  You really MUST tell the peer of all your endpoint's
> > > addresses.
> > > R
> > > "
> > 
> > Doesn't say that in RFC 2960 or anywhere else.  Here is a reason for
> > not including all the addresses:
> > 
> > I have a Signalling Gateway and an MGC, both of which have 3 interfaces:
> > two are high-speed interfaces on an Engineered network for signalling
> > applications; one is a low-speed interface which also acts as a default
> > route which is used for adminstrative tasks only (such as configuration,
> > SNMP, etc).  Although the SG can talk to the MGC on the default
> > interface, associations which are formed for signalling between the SG
> > and MGC can only use the interfaces dedicated to signalling on the
> > engineered network.
> > 
> > Now, it may be that the adminstrative interface is using DHCP on both
> > boxes.  When a DHCP lease time is renewed and a new IP address is
> > generated (e.g., on the SG), there is NO WAY that I am adding the new
> > address to the associations which are used for signalling.
> > 
> > An IPV6 example:  I have three interfaces which are connected to three
> > different service providers.  I place high-volume traffic on the two
> > service provided with which I can obtain the best rate for bulk traffic.
> > I place low-volume, low-latency traffic on the third service provider.
> > All 3 interfaces are DHCP.  I am NOT adding the low-volume service
> > provider interface to associations sending high-volume traffic.
> > 
> > I'm sorry but the INADDR_ANY approach does not work in many real-world
> > situations.
> > 
> > > >
> > > > 2)Addition and deletion of IP address shall be on Endpoint not on
> > > > Association.
> > > >
> > > > For the point 2) suppose an Endpoint E1(IPA and Port X) has multiple
> > > > associations A1, A2 and A3 with three different peers.
> > > >
> > > >     E1(IPA and PortX)                       Peer Hosts
> > > >     A1(E1)                                  IPB and Port X
> > > >     A2(E1)                                  IPC and Port X
> > > >     A3(E1)                                  IPD and Port X
> > > > As per the assumption 2) to add an IPZ at E1 SCTP has to send ASCONF chunk
> > > > on all the A1, A2 and A3 associations and it shall not consider this
> > > address
> > > > as added till it receives success from all the associations. If failure is
> > > > received is received from any of the associations then this Add-IP request
> > > > shall be discarded.
> > >
> > > Huh?  Why must it fail for all if it fails for one?  That doesn't
> > > sound too wise.  For example, if the receiver of the ASCONF does not
> > > like the new destination (perhaps it does not meet security criteria)
> > > it would not be able to reject it?
> > >
> > > Seems quite annoyed. If u are accepting ADD/DELETE IP as successfull even if
> > > ADD/DELETE IP is not successful for all the associations then again u will
> > > end up in having two instances of one transport address between two
> > > associations of same endpoint.
> > 
> > So what's the matter with that?  This is permitted by RFC 2960 in its
> > current form, with or without ADDIP.
> > 
> > --brian
> > 
> > > That is what we all had been discussing till now.
> > > ---kashif
> > >
> > > --brian
> > >
> > >
> > > >
> > > > Same procedure is to be adopted to delete an address from an endpoint.
> > > > Please comment.
> > > >
> > > > Regards
> > > > kashif
> 
> -- 
> // Jacob Heitz. Tel:510-747-2917

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 05:28:30 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA15659
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 05:28:30 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA07636
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 05:28:32 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA05684;
	Mon, 29 Apr 2002 04:50:31 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA05656
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 04:50:29 -0400 (EDT)
Received: from gorilla.mchh.siemens.de (gorilla.mchh.siemens.de [194.138.158.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA13482
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 04:50:25 -0400 (EDT)
Received: from blues.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.227])
	by gorilla.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id KAA19013;
	Mon, 29 Apr 2002 10:50:25 +0200 (MET DST)
Received: from mchh274e.demchh201e.icn.siemens.de ([139.21.200.84])
	by blues.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id KAA18919;
	Mon, 29 Apr 2002 10:49:51 +0200 (MET DST)
Received: by MCHH274E with Internet Mail Service (5.5.2653.19)
	id <268QK6R6>; Mon, 29 Apr 2002 10:50:25 +0200
Message-ID: <F92978223B01D311ACFF0008C71EE19D015DF7D2@MCHH225E>
From: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
To: "'Qiaobing Xie'" <qxie1@email.mot.com>, Ivan.Arias-Rodriguez@nokia.com
Cc: randall@stewart.chicago.il.us, s_balani@trillium.com,
        Qiaobing Xie
	 <Qiaobing.Xie@motorola.com>, mo_khan@trillium.com,
        tsvwg@ietf.org
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Mon, 29 Apr 2002 10:50:24 +0200
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Dear all,

looking at the discussion it seems that we have only one restriction 
for add-IP:

Consider the set E of all SCTP end points at a host. Partition
E = E_1 union E_2 union ... E_65535
into disjoint sets E_P of all end points having port P.

When adding an IP-address to multiple end points, all of these
end points MUST belong to different E_P.

This means that an SCTP implementation has to choose from all end points
having the same port on end point to which the add IP procedure is applied.

Another point:
Consider two associations
A1: E1 <-> E2  (The SCTP implementation of E2 is AddIP capable.)
A2: E1 <-> E3  (The SCTP implementation of E3 is not AddIP capable.)

Now you add an IP address to E1 and get E1'. So A1 is now E1' <-> E2.
Then we consider A2 also as A2 : E1' <-> E3, but the stack will never
use the added address. Is this correct usage of terminology?

Best regards
Michael

Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	Qiaobing Xie [SMTP:qxie1@email.mot.com]
> Sent:	Friday, April 26, 2002 2:45 AM
> To:	Ivan.Arias-Rodriguez@nokia.com
> Cc:	randall@stewart.chicago.il.us; s_balani@trillium.com; Qiaobing Xie; mo_khan@trillium.com; tsvwg@ietf.org
> Subject:	Re: [Tsvwg] End point and association Address conflict
> 
> Ivan.Arias-Rodriguez@nokia.com wrote:
> ....
> >         Now I have another doubt... Imagine that we have an endpoint 
> > E1 with addresses IPa, IPb and port 1. We have an association A1 
> > that only uses IPa, whose peer is IPz and port 65535. Thus the 
> > association A1 can be represented as [(IPa; Port1), (IPz; Port 65535)]. 
> > What happens now if we receive an INIT from IPz/65535 directed to IPb/1?
> > 
> >         As IPb is not being used in A1, we should answer with an 
> > INIT ACK as in a normal establishment of an association. However, 
> > when you receive an INIT chunk, you include all the addresses of the endpoint
>                                       ^^^^^^^^^^^^^
> Here your "..include all the addr..." directly violated the "under no
> circumstance can an SCTP transport address ever be present in more than
> one SCTP endpoint" rule (see last the paragraph, page 23 of the SCTP
> book).
> 
>         E1               E2
> A1 = [(IPa; Port1), (IPz; Port 65535)]
> 
>         E3                 E2
> A2 = [(IPa, IPb; Port1), (IPz; Port 65535)]
> 
> And you have IPa/Port1 present in both E1 and E3 (they are TWO different
> endpoints by definition!). This is prohibited by the SCTP association
> definition.
> 
> -Qiaobing
> 
> > in the INIT ACK chunk... But we have to be careful, because in
> > this case, when receiving the COOKIE ECHO, we have two options:
> 
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 05:30:04 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA15773
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 05:29:59 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA07683
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 05:30:01 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA06093;
	Mon, 29 Apr 2002 05:00:24 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA06054
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 05:00:21 -0400 (EDT)
Received: from gorilla.mchh.siemens.de (gorilla.mchh.siemens.de [194.138.158.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA14007
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 05:00:16 -0400 (EDT)
Received: from blues.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.227])
	by gorilla.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id LAA20045;
	Mon, 29 Apr 2002 11:00:10 +0200 (MET DST)
Received: from mchh273e.demchh201e.icn.siemens.de ([139.21.200.83])
	by blues.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id KAA21305;
	Mon, 29 Apr 2002 10:59:34 +0200 (MET DST)
Received: by MCHH273E with Internet Mail Service (5.5.2653.19)
	id <JV3Q0RDY>; Mon, 29 Apr 2002 11:00:09 +0200
Message-ID: <F92978223B01D311ACFF0008C71EE19D015DF7D3@MCHH225E>
From: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
To: "'bidulock@openss7.org'" <bidulock@openss7.org>,
        Jacob Heitz
	 <jheitz@lucent.com>
Cc: "Khan, Mohammad" <mo_khan@trillium.com>,
        "'tsvwg@ietf.org'"
	 <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: RE: [Tsvwg] EndPoint and Association Address conflict
Date: Mon, 29 Apr 2002 11:00:04 +0200
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Brian,

looking at the definition of SCTP end point on page 13 of
RFC 2960 an SCTP endpoint can be identified by a set of 
IP-addresses and a port number.

So for me it is a pair consisiting of a non empty set of
IP-addresses and a pot number. Endpoint = ({IP1, ..., IPn}, Port).

Using that definition your example is not valid if you consider
Add-IP working on end point base. It is valid if you consider 
Add-IP working on assciation base.

Best regards
Michael

Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> Sent:	Monday, April 29, 2002 10:44 AM
> To:	Jacob Heitz
> Cc:	Khan, Mohammad; 'tsvwg@ietf.org'; Pillai, Hari
> Subject:	Re: [Tsvwg] EndPoint and Association Address conflict
> 
> Jacob,
> 
> True, in the examples I gave, the Host was one SCTP endpoint, but the
> same is true for the Peer Hosts in Mohammad's example:
> 
> > > > >     E1(IPA and PortX)                       Peer Hosts
> > > > >     A1(E1)                                  IPB and Port X
> > > > >     A2(E1)                                  IPC and Port X
> > > > >     A3(E1)                                  IPD and Port X
> 
> If one defines an Endpoint as a vector of IP addresses and a single
> SCTP port number, what is to stop IP address vectors for distinct
> associations from being disjoint or overlapping in any way the
> implementation sees fit.  How one distinguishes endpoints is rather
> arbitrary.
> 
> I don't see what is stopping this:
> 
> Before:
>      Host                                    Peer Hosts
>      A1(IPA and PortX)                       IPB and Port Y
>      A2(IPA and PortX)                       IPC and Port Y
>      A3(IPA and PortX)                       IPD and Port Y
> 
> After adding IPX, IPY, IPZ:
>      Host                                    Peer Hosts
>      A1(IPA, IPX and PortX)                  IPB and Port Y
>      A2(IPA, IPY and PortX)                  IPC and Port Y
>      A3(IPA, IPZ and PortX)                  IPD and Port Y
> 
> I could see a problem binding that way in the first place, but
> for a fully formed association what does it matter?  Why define
> the endpoint for the associations as EP1 just because they start
> out life with the same IPA and PortX?
> 
> --brian
> 
> 
> On Mon, 29 Apr 2002, Jacob Heitz wrote:
> 
> > Brian,
> > 
> > You are assuming that a host is equivalent to an SCTP endpoint.
> > That is, that if a host has 3 IPs, then any SCTP endpoint in that
> > host must have the same 3 IPs. IMO, an SCTP endpoint may be
> > constructed with only a subset of the IPs of it's host. ie,
> > this host with 3 IPs may contain an SCTP endpoint that has just
> > two of those IPs.
> > 
> > "Brian F. G. Bidulock" wrote:
> > > 
> > > Mohammad,
> > > 
> > > Please see comments below...
> > > 
> > > On Mon, 29 Apr 2002, Khan, Mohammad wrote:
> > > 
> > > > Hi Brian,
> > > >
> > > > -----Original Message-----
> > > > From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> > > > Sent: Monday, April 29, 2002 12:26 PM
> > > > To: Khan, Mohammad
> > > > Cc: 'tsvwg@ietf.org'; Pillai, Hari
> > > > Subject: Re: [Tsvwg] EndPoint and Association Address conflict
> > > >
> > > >
> > > > Khan,,
> > > >
> > > > On Sun, 28 Apr 2002, Khan, Mohammad wrote:
> > > >
> > > > > Hi All,
> > > > > I could make the following observations from the ongoing discussions about
> > > > > Endpoint and Association address conflict for SCTP.
> > > > >
> > > > > 1)INIT chunk shall contain all the addresses bound to the Endpoint.
> > > >
> > > > RFC 2960 permits non-multihoming.  It also permits the addresses in
> > > > the INIT to be a subset of the addresses available to an Endpoint.
> > > > A user process starting an association might not have permission to
> > > > all the addresses owned by a host.  How do you propose to make this> 
> > > > requirement stick?
> > > >
> > > > If you mean that when an association happens to be bound with
> > > > INADDR_ANY and an IP is added, and the host determines in some
> > > > implementation dependent fashion that a new dynamic address should
> > > > be subsequently added to this association, that it should perform
> > > > the ADDIP procedures on behalf of the association, then there might
> > > > not be a problem.  But the above seems to rule out many features of
> > > > SCTP that are already implemented and being used.
> > > >
> > > > This is what Randell mentioned in one of his comments
> > > > "
> > > > Why did you not include IPb in the INIT or INIT-ACK that setup A1. You
> > > > are supposed to tell the peer of all your addresses that are bound to
> > > > your endpoint.  You really MUST tell the peer of all your endpoint's
> > > > addresses.
> > > > R
> > > > "
> > > 
> > > Doesn't say that in RFC 2960 or anywhere else.  Here is a reason for
> > > not including all the addresses:
> > > 
> > > I have a Signalling Gateway and an MGC, both of which have 3 interfaces:
> > > two are high-speed interfaces on an Engineered network for signalling
> > > applications; one is a low-speed interface which also acts as a default
> > > route which is used for adminstrative tasks only (such as configuration,
> > > SNMP, etc).  Although the SG can talk to the MGC on the default
> > > interface, associations which are formed for signalling between the SG
> > > and MGC can only use the interfaces dedicated to signalling on the
> > > engineered network.
> > > 
> > > Now, it may be that the adminstrative interface is using DHCP on both
> > > boxes.  When a DHCP lease time is renewed and a new IP address is
> > > generated (e.g., on the SG), there is NO WAY that I am adding the new
> > > address to the associations which are used for signalling.
> > > 
> > > An IPV6 example:  I have three interfaces which are connected to three
> > > different service providers.  I place high-volume traffic on the two
> > > service provided with which I can obtain the best rate for bulk traffic.
> > > I place low-volume, low-latency traffic on the third service provider.
> > > All 3 interfaces are DHCP.  I am NOT adding the low-volume service
> > > provider interface to associations sending high-volume traffic.
> > > 
> > > I'm sorry but the INADDR_ANY approach does not work in many real-world
> > > situations.
> > > 
> > > > >
> > > > > 2)Addition and deletion of IP address shall be on Endpoint not on
> > > > > Association.
> > > > >
> > > > > For the point 2) suppose an Endpoint E1(IPA and Port X) has multiple
> > > > > associations A1, A2 and A3 with three different peers.
> > > > >
> > > > >     E1(IPA and PortX)                       Peer Hosts
> > > > >     A1(E1)                                  IPB and Port X
> > > > >     A2(E1)                                  IPC and Port X
> > > > >     A3(E1)                                  IPD and Port X
> > > > > As per the assumption 2) to add an IPZ at E1 SCTP has to send ASCONF chunk
> > > > > on all the A1, A2 and A3 associations and it shall not consider this
> > > > address
> > > > > as added till it receives success from all the associations. If failure is
> > > > > received is received from any of the associations then this Add-IP request
> > > > > shall be discarded.
> > > >
> > > > Huh?  Why must it fail for all if it fails for one?  That doesn't
> > > > sound too wise.  For example, if the receiver of the ASCONF does not
> > > > like the new destination (perhaps it does not meet security criteria)
> > > > it would not be able to reject it?
> > > >
> > > > Seems quite annoyed. If u are accepting ADD/DELETE IP as successfull even if
> > > > ADD/DELETE IP is not successful for all the associations then again u will
> > > > end up in having two instances of one transport address between two
> > > > associations of same endpoint.> 
> > > 
> >>  > So what's the matter with that?  This is permitted by RFC 2960 in its
> > > current form, with or without ADDIP.
> > > 
> > > --brian
> > > 
> > > > That is what we all had been discussing till now.
> > > > ---kashif
> > > >
> > > > --brian
> > > >
> > > >
> > > > >
> > > > > Same procedure is to be adopted to delete an address from an endpoint.
> > > > > Please comment.
> > > > >
> > > > > Regards
> > > > > kashif
> > 
> > -- 
> > // Jacob Heitz. Tel:510-747-2917
> 
> -- 
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 05:36:48 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA16251
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 05:36:47 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA08369
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 05:36:50 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA06502;
	Mon, 29 Apr 2002 05:01:47 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA06473
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 05:01:45 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-230.dallas.net [209.44.42.230])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA14095
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 05:01:38 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id EAA19625;
	Mon, 29 Apr 2002 04:00:40 -0500
Date: Mon, 29 Apr 2002 04:00:40 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
Cc: Qiaobing Xie <qxie1@email.mot.com>,
        "Balani, Sandeep" <s_balani@trillium.com>,
        "'Qiaobing Xie'" <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, randall@stewart.chicago.il.us,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
Message-ID: <20020429040040.B19082@openss7.org>
Reply-To: bidulock@openss7.org
References: <F92978223B01D311ACFF0008C71EE19D015DF7D1@MCHH225E>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <F92978223B01D311ACFF0008C71EE19D015DF7D1@MCHH225E>; from Michael.Tuexen@icn.siemens.de on Mon, Apr 29, 2002 at 10:30:11AM +0200
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Michael,

On Mon, 29 Apr 2002, Tuexen Michael wrote:

> Hi Brian,
> 
> if a host A has two IP-addresses IP' and IP'' and two associations
> from (IP', Port) to a peer (IP, Port) and (IP'', Port) to the peer
> (IP, Port) then 
> - host A has two IP-addresses
> - there are two end-points (IP', Port), (IP'', Port), each of them
>   uses only one IP address available at host A.
> This 'splits' (in your terminology) host A into two end points.
> The difference between IP addresses of the end point and IP addresses
> of the host is also the reason why it is valid to say
> - you MUST include all IP address of the end point in the INIT INIT-ACK.
> - you include a subset of all addresses of the host in the INIT INIT-ACK.

Why do you call (IPA, Port) one endpoint?  I view an endpoint as
half an association.  When two associations are formed from (IPA, Port),
there are two endpoints, not one.  When adding IPB and IPC and can
add IPB to one endpoint to form (IPB, IPA, Port) and IPC to another
to form (IPC, IPA, Port).  Why restrict this?  Particularly if that is
the way I would have bound them if IPB and IPC where available in the
first place?

I have a host that talks to two other hosts.  IPA is a default and
backup interface on this host.  IPB and IPC are dedicated high-speed
interfaces for talking to HostB and HostC on an engineered network.
Normally I would bind applications serving HostB to IPB as primary
and IPA as secondary.  I would bind applications serving HostC to IPC
as primary and IPA as secondary.  I have a default server which binds
to IPA as primary.  All three servers listen on Port X.

With IPB and IPC down, the default A-server is listening on IPA Port X.
When HostB connects, it spawns a B-server and forms the association
(IPA, Port X) to HostB.  When HostC connects, it spawns an C-server
and forms the association (IPA, Port X) to HostC.  When IPB becomes
availble, the B-server wants it added to its association.  The C-server
does not.  When IPC becomes available, the C-server wants it added to
its association.  The B-Server does not.

So, unless the listener or connecter is actually bound to INADDR_ANY,
there is no reason to force it to add IP addresses just because some
other association happens to be engaged in an association with the
same local addresses and port numbers.

--brian


> 
> Best regards
> Michael
> 
> Michael Tuexen    Tel.:   +49 89 722 47210
> Siemens AG        Fax:    +49 89 722 48212
> ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de
> 
> > -----Original Message-----
> > From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> > Sent:	Monday, April 29, 2002 8:50 AM
> > To:	Qiaobing Xie
> > Cc:	Balani, Sandeep; 'Qiaobing Xie'; Ivan.Arias-Rodriguez@nokia.com; randall@stewart.chicago.il.us; Khan, Mohammad; tsvwg@ietf.org
> > Subject:	Re: [Tsvwg] End point and association Address conflict
> > 
> > Qiaobing,
> > 
> > Multi-homing is OPTIONAL.  Are you saying that a group of
> > associations that are not multi-homed (single IP address
> > in INIT), and on different IP addresses for the same host
> > is somehow splitting the host into multiple end-points?
> > I don't see how this is different than for TCP when the
> > host has multiple IP addresses: each connection may be
> > from a different IP address...  Please explain.
> > 
> > --brian
> > 
> > On Fri, 26 Apr 2002, Qiaobing Xie wrote:
> > 
> > > 
> > > I don't think anything needs to be re-defined. We just need more
> > > clarifications. To the implementor, a good add/delete-ip API semantics
> > > might help to provide a safeguard against misuses. 
> > > 
> > > SCTP is different from TCP in that it allows a single endpoint to
> > > associate with multiple peers (the "fan-out" config). But no matter how
> > > many peers the endpoint is concurrently connected to, logically it is
> > > _still_ a single endpoint. 
> > > 
> > > With this being cleared, it should not be hard to understand that it is
> > > conceptually wrong to let a user apply add/delete-ip to only a subset of
> > > the fan-out concurrent connections on that single endpoint. Because this
> > > would in fact result in "splitting" the single endpoint into two
> > > different ones. This is NEVER the intention of add/delete-ip - the
> > > feature is meant to dynamically modify an endpoint without tearing down
> > > all the associations it currently has; it is NOT meant to split a single
> > > endpoint into two!
> > > 
> > > -Qiaobing
> > > 
> > 
> > -- 
> > Brian F. G. Bidulock    > ¦ The reasonable man adapts himself to the ¦
> > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> >                         ¦ Therefore  all  progress  depends on the ¦
> >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> > 
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Apr 29 07:09:16 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA22114
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 07:09:16 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA12600
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 07:09:17 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA10296;
	Mon, 29 Apr 2002 06:24:37 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA10267
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 06:24:34 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-230.dallas.net [209.44.42.230])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA19652
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 06:24:30 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id FAA21357;
	Mon, 29 Apr 2002 05:24:10 -0500
Date: Mon, 29 Apr 2002 05:24:10 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
Cc: Jacob Heitz <jheitz@lucent.com>, "Khan, Mohammad" <mo_khan@trillium.com>,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
Message-ID: <20020429052410.A21316@openss7.org>
Reply-To: bidulock@openss7.org
References: <F92978223B01D311ACFF0008C71EE19D015DF7D3@MCHH225E>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <F92978223B01D311ACFF0008C71EE19D015DF7D3@MCHH225E>; from Michael.Tuexen@icn.siemens.de on Mon, Apr 29, 2002 at 11:00:04AM +0200
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Tuexen,

On Mon, 29 Apr 2002, Tuexen Michael wrote:

> Brian,
> 
> looking at the definition of SCTP end point on page 13 of
> RFC 2960 an SCTP endpoint can be identified by a set of 
> IP-addresses and a port number.
> 
> So for me it is a pair consisiting of a non empty set of
> IP-addresses and a pot number. Endpoint = ({IP1, ..., IPn}, Port).
> 
> Using that definition your example is not valid if you consider
> Add-IP working on end point base. It is valid if you consider 
> Add-IP working on assciation base.

Why restrict Add-IP to an endpoint basis?

--brian

> 
> Best regards
> Michael
> 
> Michael Tuexen    Tel.:   +49 89 722 47210
> Siemens AG        Fax:    +49 89 722 48212
> ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de
> 
> > -----Original Message-----
> > From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> > Sent:	Monday, April 29, 2002 10:44 AM
> > To:	Jacob Heitz
> > Cc:	Khan, Mohammad; 'tsvwg@ietf.org'; Pillai, Hari
> > Subject:	Re: [Tsvwg] EndPoint and Association Address conflict
> > 
> > Jacob,
> > 
> > True, in the examples I gave, the Host was one SCTP endpoint, but the
> > same is true for the Peer Hosts in Mohammad's example:
> > 
> > > > > >     E1(IPA and PortX)                       Peer Hosts
> > > > > >     A1(E1)                                  IPB and Port X
> > > > > >     A2(E1)                                  IPC and Port X
> > > > > >     A3(E1)                                  IPD and Port X
> > 
> > If one defines an Endpoint as a vector of IP addresses and a single
> > SCTP port number, what is to stop IP address vectors for distinct
> > associations from being disjoint or overlapping in any way the
> > implementation sees fit.  How one distinguishes endpoints is rather
> > arbitrary.
> > 
> > I don't see what is stopping this:
> > 
> > Before:
> >      Host                                    Peer Hosts
> >      A1(IPA and PortX)                       IPB and Port Y
> >      A2(IPA and PortX)                       IPC and Port Y
> >      A3(IPA and PortX)                       IPD and Port Y
> > 
> > After adding IPX, IPY, IPZ:
> >      Host                                    Peer Hosts
> >      A1(IPA, IPX and PortX)                  IPB and Port Y
> >      A2(IPA, IPY and PortX)                  IPC and Port Y
> >      A3(IPA, IPZ and PortX)                  IPD and Port Y
> > 
> > I could see a problem binding that way in the first place, but
> > for a fully formed association what does it matter?  Why define
> > the endpoint for the associations as EP1 just because they start
> > out life with the same IPA and PortX?
> > 
> > --brian
> > 
> > 
> > On Mon, 29 Apr 2002, Jacob Heitz wrote:
> > 
> > > Brian,
> > > 
> > > You are assuming that a host is equivalent to an SCTP endpoint.
> > > That is, that if a host has 3 IPs, then any SCTP endpoint in that
> > > host must have the same 3 IPs. IMO, an SCTP endpoint may be
> > > constructed with only a subset of the IPs of it's host. ie,
> > > this host with 3 IPs may contain an SCTP endpoint that has just
> > > two of those IPs.
> > > 
> > > "Brian F. G. Bidulock" wrote:
> > > > 
> > > > Mohammad,
> > > > 
> > > > Please see comments below...
> > > > 
> > > > On Mon, 29 Apr 2002, Khan, Mohammad wrote:
> > > > 
> > > > > Hi Brian,
> > > > >
> > > > > -----Original Message-----
> > > > > From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> > > > > Sent: Monday, April 29, 2002 12:26 PM
> > > > > To: Khan, Mohammad
> > > > > Cc: 'tsvwg@ietf.org'; Pillai, Hari
> > > > > Subject: Re: [Tsvwg] EndPoint and Association Address conflict
> > > > >
> > > > >
> > > > > Khan,,
> > > > >
> > > > > On Sun, 28 Apr 2002, Khan, Mohammad wrote:
> > > > >
> > > > > > Hi All,
> > > > > > I could make the following observations from the ongoing discussions about
> > > > > > Endpoint and Association address conflict for SCTP.
> > > > > >
> > > > > > 1)INIT chunk shall contain all the addresses bound to the Endpoint.
> > > > >
> > > > > RFC 2960 permits non-multihoming.  It also permits the addresses in
> > > > > the INIT to be a subset of the addresses available to an Endpoint.
> > > > > A user process starting an association might not have permission to
> > > > > all the addresses owned by a host.  How do you propose to make this> 
> > > > > requirement stick?
> > > > >
> > > > > If you mean that when an association happens to be bound with
> > > > > INADDR_ANY and an IP is added, and the host determines in some
> > > > > implementation dependent fashion that a new dynamic address should
> > > > > be subsequently added to this association, that it should perform
> > > > > the ADDIP procedures on behalf of the association, then there might
> > > > > not be a problem.  But the above seems to rule out many features of
> > > > > SCTP that are already implemented and being used.
> > > > >
> > > > > This is what Randell mentioned in one of his comments
> > > > > "
> > > > > Why did you not include IPb in the INIT or INIT-ACK that setup A1. You
> > > > > are supposed to tell the peer of all your addresses that are bound to
> > > > > your endpoint.  You really MUST tell the peer of all your endpoint's
> > > > > addresses.
> > > > > R
> > > > > "
> > > > 
> > > > Doesn't say that in RFC 2960 or anywhere else.  Here is a reason for
> > > > not including all the addresses:
> > > > 
> > > > I have a Signalling Gateway and an MGC, both of which have 3 interfaces:
> > > > two are high-speed interfaces on an Engineered network for signalling
> > > > applications; one is a low-speed interface which also acts as a default
> > > > route which is used for adminstrative tasks only (such as configuration,
> > > > SNMP, etc).  Although the SG can talk to the MGC on the default
> > > > interface, associations which are formed for signalling between the SG
> > > > and MGC can only use the interfaces dedicated to signalling on the
> > > > engineered network.
> > > > 
> > > > Now, it may be that the adminstrative interface is using DHCP on both
> > > > boxes.  When a DHCP lease time is renewed and a new IP address is
> > > > generated (e.g., on the SG), there is NO WAY that I am adding the new
> > > > address to the associations which are used for signalling.
> > > > 
> > > > An IPV6 example:  I have three interfaces which are connected to three
> > > > different service providers.  I place high-volume traffic on the two
> > > > service provided with which I can obtain the best rate for bulk traffic.
> > > > I place low-volume, low-latency traffic on the third service provider.
> > > > All 3 interfaces are DHCP.  I am NOT adding the low-volume service
> > > > provider interface to associations sending high-volume traffic.
> > > > 
> > > > I'm sorry but the INADDR_ANY approach does not work in many real-world
> > > > situations.
> > > > 
> > > > > >
> > > > > > 2)Addition and deletion of IP address shall be on Endpoint not on
> > > > > > Association.
> > > > > >
> > > > > > For the point 2) suppose an Endpoint E1(IPA and Port X) has multiple
> > > > > > associations A1, A2 and A3 with three different peers.
> > > > > >
> > > > > >     E1(IPA and PortX)                       Peer Hosts
> > > > > >     A1(E1)                                  IPB and Port X
> > > > > >     A2(E1)                                  IPC and Port X
> > > > > >     A3(E1)                                  IPD and Port X
> > > > > > As per the assumption 2) to add an IPZ at E1 SCTP has to send ASCONF chunk
> > > > > > on all the A1, A2 and A3 associations and it shall not consider this
> > > > > address
> > > > > > as added till it receives success from all the associations. If failure is
> > > > > > received is received from any of the associations then this Add-IP request
> > > > > > shall be discarded.
> > > > >
> > > > > Huh?  Why must it fail for all if it fails for one?  That doesn't
> > > > > sound too wise.  For example, if the receiver of the ASCONF does not
> > > > > like the new destination (perhaps it does not meet security criteria)
> > > > > it would not be able to reject it?
> > > > >
> > > > > Seems quite annoyed. If u are accepting ADD/DELETE IP as successfull even if
> > > > > ADD/DELETE IP is not successful for all the associations then again u will
> > > > > end up in having two instances of one transport address between two
> > > > > associations of same endpoint.> 
> > > > 
> > >>  > So what's the matter with that?  This is permitted by RFC 2960 in its
> > > > current form, with or without ADDIP.
> > > > 
> > > > --brian
> > > > 
> > > > > That is what we all had been discussing till now.
> > > > > ---kashif
> > > > >
> > > > > --brian
> > > > >
> > > > >
> > > > > >
> > > > > > Same procedure is to be adopted to delete an address from an endpoint.
> > > > > > Please comment.
> > > > > >
> > > > > > Regards
> > > > > > kashif
> > > 
> > > -- 
> > > // Jacob Heitz. Tel:510-747-2917
> > 
> > -- 
> > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> >                         ¦ Therefore  all  progress  depends on the ¦
> >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> > 
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Apr 29 07:58:59 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA26009
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 07:58:59 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA14778
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 07:59:00 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA13663;
	Mon, 29 Apr 2002 07:32:15 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA13581
	for <tsvwg@optimus.ietf.org>; Mon, 29 Apr 2002 07:32:09 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-230.dallas.net [209.44.42.230])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA23254
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 07:32:04 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id GAA22225;
	Mon, 29 Apr 2002 06:31:27 -0500
Date: Mon, 29 Apr 2002 06:31:27 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
Cc: Qiaobing Xie <qxie1@email.mot.com>,
        "Balani, Sandeep" <s_balani@trillium.com>,
        "'Qiaobing Xie'" <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, randall@stewart.chicago.il.us,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
Message-ID: <20020429063127.B21316@openss7.org>
Reply-To: bidulock@openss7.org
References: <F92978223B01D311ACFF0008C71EE19D015DF7D7@MCHH225E>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <F92978223B01D311ACFF0008C71EE19D015DF7D7@MCHH225E>; from Michael.Tuexen@icn.siemens.de on Mon, Apr 29, 2002 at 12:53:33PM +0200
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Michael,

Please see comments below...

On Mon, 29 Apr 2002, Tuexen Michael wrote:

> > On Mon, 29 Apr 2002, Tuexen Michael wrote:
> > 
> > > Hi Brian,
> > > 
> > > if a host A has two IP-addresses IP' and IP'' and two associations
> > > from (IP', Port) to a peer (IP, Port) and (IP'', Port) to the peer
> > > (IP, Port) then 
> > > - host A has two IP-addresses
> > > - there are two end-points (IP', Port), (IP'', Port), each of them
> > >   uses only one IP address available at host A.
> > > This 'splits' (in your terminology) host A into two end points.
> > > The difference between IP addresses of the end point and IP addresses
> > > of the host is also the reason why it is valid to say
> > > - you MUST include all IP address of the end point in the INIT INIT-ACK.
> > > - you include a subset of all addresses of the host in the INIT INIT-ACK.
> > 
> > Why do you call (IPA, Port) one endpoint?  I view an endpoint as
> > half an association.  When two associations are formed from (IPA, Port),
> > there are two endpoints, not one.  When adding IPB and IPC and can
> > add IPB to one endpoint to form (IPB, IPA, Port) and IPC to another
> > to form (IPC, IPA, Port).  Why restrict this?  Particularly if that is
> > the way I would have bound them if IPB and IPC where available in the
> > first place?
> 
> [Tuexen Michael]  I think we are not using the same definition of end
> point. I'm using an abstract one based on RFC 2960. 
> Def.: An SCTP end point is a pair of a set of IP addresses and a port
>       number.

We all agree that an endpoint is a set of IP addresses and a port.  This
does not say that an endpoint is a unique set of IP addresses and a
port.  The cardinality of endpoints is not defined.

> This means, considering a server, I have only one end point for the
> server.

Your interpretation.  I have given an example where more than one server
can service the same endpoint.  This is typical and true even for TCP
(where multiple HTTP servers, for example, service the same IP/Port).

> Mulitple clients can conncect to the server.  These associations share
> the same end point.

Depending on how one views the cardinality of endpoints, these
associations can be viewed as connecting to different endpoints (even
though those endpoints were orginally bound to the same IP...IP/Port,
that is they had the same name).

> 
> This definition is in tune with the sentence "Two SCTP end points MUST
> not have more than one SCTP association between them at any time
> given" RFC 2960, page 13, defintion of SCTP association.

This gives no clues as to the cardinality of endpoints.  It only states
the cardinality of associations.

> 
> Using your definition as an end point is "half of an association" The
> sentence cited abover would be trivial.

The sentence above is what makes an endpoint half an association...

> 
> That is why I'm still thinking that my notion of SCTP end point is the
> one the authors of RFC 2960 had in mind.

Endpoint has had far more before RFC 2960.  An endpoint for
communication is the full duplex communications endpoint which occurs at
either end of the association.  When we bind and endpoint (give it a
name) we associate the IP...IP/Port with the endpoint.  When we connect
an association, we form a unique connection to another endpoint.  That
endpoint also has an name: IP...IP/Port, but we only contact it using
one IP address.

Just as with TCP, if we make an active connection from an endpoint only
(that is, we are not listening) then multiple sockets/streams can be
bound (named) with the same IP...IP/Port combination.  In fact, with
dynamic port allocation, IP/Port combinations are reused on TCP every
day between multiple connecting TCP endpoints (sockets/streams) that
have no clue that the magic port number `0' has yeilded the same IP/Port
combination as another TCP endpoint (socket/stream).

This is also the case of SCTP.  I have two sockets which I have bound
to IPA/IPB/Port`0'.  I connect from one socket to Host B and the
implementation assigns me port 10000.  I connect from the other socket
to Host C and the implementation also assigns me port 10000  (knowing
that there is no conflict because one IPA/IPB/Port10000 socket is
connected to HostB and the other IPA/IPB/Port10000 socket is connected
to HostC).  Now how can you call these the same endpoint?

--brian

> 
> Best regards
> Michael
> 

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Apr 29 08:24:42 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA28352
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 08:24:41 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA16666
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 08:24:43 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA14630;
	Mon, 29 Apr 2002 07:55:59 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA14593
	for <tsvwg@optimus.ietf.org>; Mon, 29 Apr 2002 07:55:55 -0400 (EDT)
Received: from gorilla.mchh.siemens.de (gorilla.mchh.siemens.de [194.138.158.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA25457
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 07:55:51 -0400 (EDT)
Received: from moody.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.226])
	by gorilla.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id NAA05133;
	Mon, 29 Apr 2002 13:55:43 +0200 (MET DST)
Received: from mchh246e.demchh201e.icn.siemens.de ([139.21.200.56])
	by moody.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id NAA10089;
	Mon, 29 Apr 2002 13:55:45 +0200 (MET DST)
Received: by MCHH246E with Internet Mail Service (5.5.2653.19)
	id <JT90NXPZ>; Mon, 29 Apr 2002 13:55:45 +0200
Message-ID: <F92978223B01D311ACFF0008C71EE19D015DF7DB@MCHH225E>
From: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
To: "'bidulock@openss7.org'" <bidulock@openss7.org>,
        Tuexen Michael
	 <Michael.Tuexen@icn.siemens.de>
Cc: Jacob Heitz <jheitz@lucent.com>, "Khan, Mohammad"
	 <mo_khan@trillium.com>,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: RE: [Tsvwg] EndPoint and Association Address conflict
Date: Mon, 29 Apr 2002 13:55:37 +0200
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Brian,

see my comments below.

Best regards
Michael

Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> Sent:	Monday, April 29, 2002 12:24 PM
> To:	Tuexen Michael
> Cc:	Jacob Heitz; Khan, Mohammad; 'tsvwg@ietf.org'; Pillai, Hari
> Subject:	Re: [Tsvwg] EndPoint and Association Address conflict
> 
> Tuexen,
> 
> On Mon, 29 Apr 2002, Tuexen Michael wrote:
> 
> > Brian,
> > 
> > looking at the definition of SCTP end point on page 13 of
> > RFC 2960 an SCTP endpoint can be identified by a set of 
> > IP-addresses and a port number.
> > 
> > So for me it is a pair consisiting of a non empty set of
> > IP-addresses and a pot number. Endpoint = ({IP1, ..., IPn}, Port).
> > 
> > Using that definition your example is not valid if you consider
> > Add-IP working on end point base. It is valid if you consider 
> > Add-IP working on assciation base.
> 
> Why restrict Add-IP to an endpoint basis?
> 
	[Tuexen Michael]  Consider E1 = ({IP'}, Port), E2 = ({IP'''}, Port),
	E3 = ({IP''''}, Port) with pairwise different IP', IP''', IP'''' and two associations
	A1: E1 <-> E2
	A2: E1 <-> E3

	Now if we add IP'' to the association A1 we get
	A1 : E1' <-> E2
	with E1' = ({IP', IP''}, Port).
	E1' and E2 share one transport address which is in violation to RFC 2960, page 13.

	Please note that an association has two SCTP end points and these end points
	have the IP addresses. So changing IP addresses means modifying the end points.

	Best regards
	Michael  
> --brian
> 
> > 
> > Best regards
> > Michael
> > 
> > Michael Tuexen    Tel.:   +49 89 722 47210
> > Siemens AG        Fax:    +49 89 722 48212
> > ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de
> > 
> > > -----Original Message-----
> > > From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> > > Sent:	Monday, April 29, 2002 10:44 AM
> > > To:	Jacob Heitz
> > > Cc:	Khan, Mohammad; 'tsvwg@ietf.org'; Pillai, Hari
> > > Subject:	Re: [Tsvwg] EndPoint and Association Address conflict
> > > 
> > > Jacob,
> > > 
> > > True, in the examples I gave, the Host was one SCTP endpoint, but the
> > > same is true for the Peer Hosts in Mohammad's example:
> > > 
> > > > > > >     E1(IPA and PortX)                       Peer Hosts
> > > > > > >     A1(E1)                                  IPB and Port X
> > > > > > >     A2(E1)                                  IPC and Port X
> > > > > > >     A3(E1)                                  IPD and Port X
> > > 
> > > If one defines an Endpoint as a vector of IP addresses and a single
> > > SCTP port number, what is to stop IP address vectors for distinct
> > > associations from being disjoint or overlapping in any way the
> > > implementation sees fit.  How one distinguishes endpoints is rather
> > > arbitrary.
> > > 
> > > I don't see what is stopping this:
> > > 
> > > Before:
> > >      Host                                    Peer Hosts
> > >      A1(IPA and PortX)                       IPB and Port Y
> > >      A2(IPA and PortX)                       IPC and Port Y
> > >      A3(IPA and PortX)                       IPD and Port Y
> > > 
> > > After adding IPX, IPY, IPZ:
> > >      Host                                    Peer Hosts
> > >      A1(IPA, IPX and PortX)                  IPB and Port Y
> > >      A2(IPA, IPY and PortX)                  IPC and Port Y
> > >      A3(IPA, IPZ and PortX)                  IPD and Port Y
> > > 
> > > I could see a problem binding that way in the first place, but
> > > for a fully formed association what does it matter?  Why define
> > > the endpoint for the associations as EP1 just because they start
> > > out life with the same IPA and PortX?
> > > 
> > > --brian
> > > 
> > > 
> > > On Mon, 29 Apr 2002, Jacob Heitz wrote:> 
> > > 
> > > > Brian,
> > > > 
> > > > You are assuming that a host is equivalent to an SCTP endpoint.
> > > > That is, that if a host has 3 IPs, then any SCTP endpoint in that
> > > > host must have the same 3 IPs. IMO, an SCTP endpoint may be
> > > > constructed with only a subset of the IPs of it's host. ie,
> > > > this host with 3 IPs may contain an SCTP endpoint that has just
> > > > two of those IPs.
> > > > 
> > > > "Brian F. G. Bidulock" wrote:
> > > > > 
> > > > > Mohammad,
> > > > > 
> > > > > Please see comments below...
> > > > > 
> > > > > On Mon, 29 Apr 2002, Khan, Mohammad wrote:
> > > > > 
> > > > > > Hi Brian,
> > > > > >
> > > > > > -----Original Message-----
> > > > > > From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> > > > > > Sent: Monday, April 29, 2002 12:26 PM
> > > > > > To: Khan, Mohammad
> > > > > > Cc: 'tsvwg@ietf.org'; Pillai, Hari
> > > > > > Subject: Re: [Tsvwg] EndPoint and Association Address conflict
> > > > > >
> > > > > >
> > > > > > Khan,,
> > > > > >
> > > > > > On Sun, 28 Apr 2002, Khan, Mohammad wrote:
> > > > > >
> > > > > > > Hi All,
> > > > > > > I could make the following observations from the ongoing discussions about
> > > > > > > Endpoint and Association address conflict for SCTP.
> > > > > > >
> > > > > > > 1)INIT chunk shall contain all the addresses bound to the Endpoint.
> > > > > >
> > > > > > RFC 2960 permits non-multihoming.  It also permits the addresses in
> > > > > > the INIT to be a subset of the addresses available to an Endpoint.
> > > > > > A user process starting an association might not have permission to
> > > > > > all the addresses owned by a host.  How do you propose to make this> 
> > > > > > requirement stick?
> > > > > >
> > > > > > If you mean that when an association happens to be bound with
> > > > > > INADDR_ANY and an IP is added, and the host determines in some
> > > > > > implementation dependent fashion that a new dynamic address should
> > > > > > be subsequently added to this association, that it should perform
> > > > > > the ADDIP procedures on behalf of the association, then there might
> > > > > > not be a problem.  But the above seems to rule out many features of
> > > > > > SCTP that are already implemented and being used.
> > > > > >
> > > > > > This is what Randell mentioned in one of his comments
> > > > > > "
> > > > > > Why did you not include IPb in the INIT or INIT-ACK that setup A1. You
> > > > > > are supposed to tell the peer of all your addresses that are bound to
> > > > > > your endpoint.  You really MUST tell the peer of all your endpoint's
> > > > > > addresses.
> > > > > > R
> > > > > > "
> > > > > 
> > > > > Doesn't say that in RFC 2960 or anywhere else.  Here is a reason for
> > > > > not including all the addresses:
> > > > > 
> > > > > I have a Signalling Gateway and an MGC, both of which have 3 interfaces:
> > > > > two are high-speed interfaces on an Engineered network for signalling
> > > > > applications; one is a low-speed interface which also acts as a default
> > > > > route which is used for adminstrative tasks only (such as configuration,
> > > > > SNMP, etc).  Although the SG can talk to the MGC on the default
> > > > > interface, associations which are formed for signalling between the SG
> > > > > and MGC can only use the interfaces dedicated to signalling on the
> > > > > engineered network.
> > > > > 
> > > > > Now, it may be that the adminstrative interface is using DHCP on both
> > > > > boxes.  When a DHCP lease time is renewed and a new IP address is
> > > > > generated (e.g., on the SG), there is NO WAY that I am adding the new
> > > > > address to the associations which are used for signalling.
> > > > > 
> > > > > An IPV6 example:  I have three interfaces which are connected to three
> > > > > different service providers.  I place high-volume traffic on the two
> > > > > service provided with which I can obtain the best rate for bulk traffic.> 
> > > > > I pla> ce low-volume, low-latency traffic on the third service provider.
> > > > > All 3 interfaces are DHCP.  I am NOT adding the low-volume service
> > > > > provider interface to associations sending high-volume traffic.
> > > > > 
> > > > > I'm sorry but the INADDR_ANY approach does not work in many real-world
> > > > > situations.
> > > > > 
> > > > > > >
> > > > > > > 2)Addition and deletion of IP address shall be on Endpoint not on
> > > > > > > Association.
> > > > > > >
> > > > > > > For the point 2) suppose an Endpoint E1(IPA and Port X) has multiple
> > > > > > > associations A1, A2 and A3 with three different peers.
> > > > > > >
> > > > > > >     E1(IPA and PortX)                       Peer Hosts
> > > > > > >     A1(E1)                                  IPB and Port X
> > > > > > >     A2(E1)                                  IPC and Port X
> > > > > > >     A3(E1)                                  IPD and Port X
> > > > > > > As per the assumption 2) to add an IPZ at E1 SCTP has to send ASCONF chunk
> > > > > > > on all the A1, A2 and A3 associations and it shall not consider this
> > > > > > address
> > > > > > > as added till it receives success from all the associations. If failure is
> > > > > > > received is received from any of the associations then this Add-IP request
> > > > > > > shall be discarded.
> > > > > >
> > > > > > Huh?  Why must it fail for all if it fails for one?  That doesn't
> > > > > > sound too wise.  For example, if the receiver of the ASCONF does not
> > > > > > like the new destination (perhaps it does not meet security criteria)
> > > > > > it would not be able to reject it?
> > > > > >
> > > > > > Seems quite annoyed. If u are accepting ADD/DELETE IP as successfull even if
> > > > > > ADD/DELETE IP is not successful for all the associations then again u will
> > > > > > end up in having two instances of one transport address between two
> > > > > > associations of same endpoint.> 
> > > > > 
> > > >>  > So what's the matter with that?  This is permitted by RFC 2960 in its
> > > > > current form, with or without ADDIP.
> > > > > 
> > > > > --brian
> > > > > 
> > > > > > That is what we all had been discussing till now.
> > > > > > ---kashif
> > > > > >
> > > > > > --brian
> > > > > >
> > > > > >
> > > > > > >
> > > > > > > Same procedure is to be adopted to delete an address from an endpoint.
> > > > > > > Please comment.
> > > > > > >
> > > > > > > Regards
> > > > > > > kashif
> > > > 
> > > > -- 
> > > > // Jacob Heitz. Tel:510-747-2917
> > > 
> > > -- 
> > > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> > >                         ¦ Therefore  all  progress  depends on the ¦
> > >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> > > 
> > > _______________________________________________
> > > tsvwg mailing list
> > > tsvwg@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> -- 
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Apr 29 08:37:14 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA29503
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 08:37:14 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA17809
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 08:37:16 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA16209;
	Mon, 29 Apr 2002 08:15:14 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA16179
	for <tsvwg@optimus.ietf.org>; Mon, 29 Apr 2002 08:15:11 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-230.dallas.net [209.44.42.230])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA27489
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 08:15:08 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id HAA22809;
	Mon, 29 Apr 2002 07:14:52 -0500
Date: Mon, 29 Apr 2002 07:14:52 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
Cc: Jacob Heitz <jheitz@lucent.com>, "Khan, Mohammad" <mo_khan@trillium.com>,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
Message-ID: <20020429071452.C21316@openss7.org>
Reply-To: bidulock@openss7.org
References: <F92978223B01D311ACFF0008C71EE19D015DF7DB@MCHH225E>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <F92978223B01D311ACFF0008C71EE19D015DF7DB@MCHH225E>; from Michael.Tuexen@icn.siemens.de on Mon, Apr 29, 2002 at 01:55:37PM +0200
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Tuexen,

On Mon, 29 Apr 2002, Tuexen Michael wrote:
> > 
> > Why restrict Add-IP to an endpoint basis?
> > 
> 	[Tuexen Michael]  Consider E1 = ({IP'}, Port), E2 = ({IP'''}, Port),
> 	E3 = ({IP''''}, Port) with pairwise different IP', IP''', IP'''' and two associations
> 	A1: E1 <-> E2
> 	A2: E1 <-> E3
> 
> 	Now if we add IP'' to the association A1 we get
> 	A1 : E1' <-> E2
> 	with E1' = ({IP', IP''}, Port).
> 	E1' and E2 share one transport address which is in violation to RFC
> 	2960, page 13.

There is only one MUST on page 13: "Two SCTP endpoints MUST NOT have more
than one SCTP association between them at any given time."

How does this violate anything on page 13?

E1' and E2 only have one association between them.  E1 and E3 only have one
association between them.  Everthing appears ok as far as page 13 is
concerned.

Did you mean another page?

Or, did you mean the statement: "A transport address used by an
SCTP endpoint must not be used by another SCTP endpoint.  In other
words, a transport address is unique to an SCTP endpoint."  ...which
is absurd, particularly as it is done in TCP all the time, and when
you consider that non-multihomed SCTP associations need to reuse
dynamic port numbers for connection to other hosts.

--brian

> 
> 	Please note that an association has two SCTP end points and these end
> 	points have the IP addresses. So changing IP addresses means modifying
> 	the end points.
> 
> 	Best regards
> 	Michael  

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Apr 29 08:39:02 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA29669
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 08:38:48 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA17900
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 08:38:50 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA16305;
	Mon, 29 Apr 2002 08:17:04 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA16272
	for <tsvwg@optimus.ietf.org>; Mon, 29 Apr 2002 08:17:01 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-230.dallas.net [209.44.42.230])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA27653
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 08:16:57 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id HAA22838;
	Mon, 29 Apr 2002 07:16:25 -0500
Date: Mon, 29 Apr 2002 07:16:25 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
Cc: Qiaobing Xie <qxie1@email.mot.com>,
        "Balani, Sandeep" <s_balani@trillium.com>,
        "'Qiaobing Xie'" <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, randall@stewart.chicago.il.us,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
Message-ID: <20020429071625.D21316@openss7.org>
Reply-To: bidulock@openss7.org
References: <F92978223B01D311ACFF0008C71EE19D015DF7D9@MCHH225E>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <F92978223B01D311ACFF0008C71EE19D015DF7D9@MCHH225E>; from Michael.Tuexen@icn.siemens.de on Mon, Apr 29, 2002 at 01:41:31PM +0200
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Michael,

You may as well let the RFC talk for itself, then.

--brian

On Mon, 29 Apr 2002, Tuexen Michael wrote:

> Brian,
> 
> I'm talking ONLY about an 'SCTP endpoint', as being defined in
> RFC 2960. I agree, that the term endpoint is used with a lot
> of meanings in natural language. I'm using it not implying any
> other meaning (for example for TCP or so), it is just a term
> defined in RFC 2960. 
> 
> I do understand your examples and I only say that the notion 
> defined in RFC 2960 does not apply to them in a way you are using
> them. 
> 
> Maybe you think that is the wrong notion, but I think that is the 
> one used in RFC 2960 and the extension IDs. Since it is necessary
> to use the same one, Randy or Qiaobing could state which one they
> had in mind when writing the RFC. Then we can all use the same.
> 
> Best regards
> Michael
> Michael Tuexen    Tel.:   +49 89 722 47210
> Siemens AG        Fax:    +49 89 722 48212
> ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de
> 
> > -----Original Message-----
> > From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> > Sent:	Monday, April 29, 2002 1:31 PM
> > To:	Tuexen Michael
> > Cc:	Qiaobing Xie; Balani, Sandeep; 'Qiaobing Xie'; Ivan.Arias-Rodriguez@nokia.com; randall@stewart.chicago.il.us; Khan, Mohammad; tsvwg@ietf.org
> > Subject:	Re: [Tsvwg] End point and association Address conflict
> > 
> > Michael,
> > 
> > Please see comments below...
> > 
> > On Mon, 29 Apr 2002, Tuexen Michael wrote:
> > 
> > > > On Mon, 29 Apr 2002, Tuexen Michael wrote:
> > > > 
> > > > > Hi Brian,
> > > > > 
> > > > > if a host A has two IP-addresses IP' and IP'' and two associations
> > > > > from (IP', Port) to a peer (IP, Port) and (IP'', Port) to the peer
> > > > > (IP, Port) then 
> > > > > - host A has two IP-addresses
> > > > > - there are two end-points (IP', Port), (IP'', Port), each of them
> > > > >   uses only one IP address available at host A.
> > > > > This 'splits' (in your terminology) host A into two end points.
> > > > > The difference between IP addresses of the end point and IP addresses
> > > > > of the host is also the reason why it is valid to say
> > > > > - you MUST include all IP address of the end point in the INIT INIT-ACK.
> > > > > - you include a subset of all addresses of the host in the INIT INIT-ACK.
> > > > 
> > > > Why do you call (IPA, Port) one endpoint?  I view an endpoint as
> > > > half an association.  When two associations are formed from (IPA, Port),
> > > > there are two endpoints, not one.  When adding IPB and IPC and can
> > > > add IPB to one endpoint to form (IPB, IPA, Port) and IPC to another
> > > > to form (IPC, IPA, Port).  Why restrict this?  Particularly if that is
> > > > the way I would have bound them if IPB and IPC where available in the
> > > > first place?
> > > 
> > > [Tuexen Michael]  I think we are not using the same definition of end
> > > point. I'm using an abstract one based on RFC 2960. 
> > > Def.: An SCTP end point is a pair of a set of IP addresses and a port
> > >       number.
> > 
> > We all agree that an endpoint is a set of IP addresses and a port.  This
> > does not say that an endpoint is a unique set of IP addresses and a
> > port.  The cardinality of endpoints is not defined.
> > 
> > > This means, considering a server, I have only one end point for the
> > > server.
> > 
> > Your interpretation.  I have given an example where more than one server
> > can service the same endpoint.  This is typical and true even for TCP
> > (where multiple HTTP servers, for example, service the same IP/Port).
> > 
> > > Mulitple clients can conncect to the server.  These associations share
> > > the same end point.
> > 
> > Depending on how one views the cardinality of endpoints, these
> > associations can be viewed as connecting to different endpoints (even
> > though those endpoints were orginally bound to the same IP...IP/Port,
> > that is they had the same name).
> > 
> > > 
> > > This definition is in tune with the sentence "Two SCTP end points MUST
> > > not have more than one SCTP association between them at any time
> > > given" RFC 2960, page 13, defintion of SCTP association.
> > 
> > This gives no clues as to the cardinality of endpoints.  It only states
> > the cardinality of associations.
> > 
> > > 
> > > Using your definition as an end point is "half of an association" The
> > > sentence cited abover would be trivial.
> > 
> > The sentence above is what makes an endpoint half an association...
> > 
> > > 
> > > That is why I'm still thinking that my notion of SCTP end point is the
> > > one the authors of RFC 2960 had in mind.
> > 
> > Endpoint has had far more before RFC 2960.  An endpoint for
> > communication is the full duplex communications endpoint which occurs at
> > either end of the association.  When we bind and endpoint (give it a
> > name) we associate the IP...IP/Port with the endpoint.  When we connect
> > an association, we form a unique connection to another endpoint.  That
> > endpoint also has an name: IP...IP/Port, but we only contact it using
> > one IP address.
> > 
> > Just as with TCP, if we make an active connection from an endpoint only
> > (that is, we are not listening) then multiple sockets/streams can be
> > bound (named) with the same IP...IP/Port combination.  In fact, with
> > dynamic port allocation, IP/Port combinations are reused on TCP every
> > day between multiple connecting TCP endpoints (sockets/streams) that
> > have no clue that the magic port number `0' has yeilded the same IP/Port
> > combination as another TCP endpoint (socket/stream).
> > 
> > This is also the case of SCTP.  I have two sockets which I have bound
> > to IPA/IPB/Port`0'.  I connect from one socket to Host B and the
> > implementation assigns me port 10000.  I connect from the other socket
> > to Host C and the implementation also assigns me port 10000  (knowing
> > that there is no conflict because one IPA/IPB/Port10000 socket is
> > connected to HostB and the other IPA/IPB/Port10000 socket is connected
> > to HostC).  Now how can you call these the same endpoint?
> > 
> > --brian
> > 
> > > 
> > > Best regards
> > > Michael
> > > 
> > 
> > -- 
> > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> >                         ¦ Therefore  all  progress  depends on the ¦
> >                         ¦ unreasonable man. -- George Bernard Shaw ¦

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Apr 29 08:42:37 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA29942
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 08:42:36 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA17992
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 08:42:37 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA16456;
	Mon, 29 Apr 2002 08:20:25 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA16425
	for <tsvwg@optimus.ietf.org>; Mon, 29 Apr 2002 08:20:21 -0400 (EDT)
Received: from beamer.mchh.siemens.de (beamer.mchh.siemens.de [194.138.158.163])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA27929
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 08:20:19 -0400 (EDT)
Received: from blues.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.227])
	by beamer.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id OAA24592;
	Mon, 29 Apr 2002 14:20:02 +0200 (MET DST)
Received: from mchh273e.demchh201e.icn.siemens.de ([139.21.200.83])
	by blues.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id OAA02827;
	Mon, 29 Apr 2002 14:19:35 +0200 (MET DST)
Received: by MCHH273E with Internet Mail Service (5.5.2653.19)
	id <JV3Q0ZSS>; Mon, 29 Apr 2002 14:20:10 +0200
Message-ID: <F92978223B01D311ACFF0008C71EE19D015DF7DC@MCHH225E>
From: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
To: "'bidulock@openss7.org'" <bidulock@openss7.org>
Cc: Jacob Heitz <jheitz@lucent.com>, "Khan, Mohammad"
	 <mo_khan@trillium.com>,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: RE: [Tsvwg] EndPoint and Association Address conflict
Date: Mon, 29 Apr 2002 14:20:03 +0200
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id IAA16426
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Brian,

see my comments below.

Best regards
Michael

Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> Sent:	Monday, April 29, 2002 2:15 PM
> To:	Tuexen Michael
> Cc:	Jacob Heitz; Khan, Mohammad; 'tsvwg@ietf.org'; Pillai, Hari
> Subject:	Re: [Tsvwg] EndPoint and Association Address conflict
> 
> Tuexen,
> 
> On Mon, 29 Apr 2002, Tuexen Michael wrote:
> > > 
> > > Why restrict Add-IP to an endpoint basis?
> > > 
> > 	[Tuexen Michael]  Consider E1 = ({IP'}, Port), E2 = ({IP'''}, Port),
> > 	E3 = ({IP''''}, Port) with pairwise different IP', IP''', IP'''' and two associations
> > 	A1: E1 <-> E2
> > 	A2: E1 <-> E3
> > 
> > 	Now if we add IP'' to the association A1 we get
> > 	A1 : E1' <-> E2
> > 	with E1' = ({IP', IP''}, Port).
> > 	E1' and E2 share one transport address which is in violation to RFC
> > 	2960, page 13.
> 
> There is only one MUST on page 13: "Two SCTP endpoints MUST NOT have more
> than one SCTP association between them at any given time."
> 
> How does this violate anything on page 13?
	[Tuexen Michael]  I did not mean this sentence. 
> E1' and E2 only have one association between them.  E1 and E3 only have one
> association between them.  Everthing appears ok as far as page 13 is
> concerned.
> 
> Did you mean another page?
> 
> Or, did you mean the statement: "A transport address used by an
> SCTP endpoint must not be used by another SCTP endpoint.  In other
> words, a transport address is unique to an SCTP endpoint."  ...which
> is absurd, particularly as it is done in TCP all the time, and when
> you consider that non-multihomed SCTP associations need to reuse
> dynamic port numbers for connection to other hosts.
	[Tuexen Michael]  I mean this sentence. This sentence
	is NOT absurd, at least if you use the definition of
	SCTP end-point I have given earlier. This is one of the reasons
	I'm quite sure that the RFC authors had the defintion I gave
	in mind.

> --brian
> 
> > 
> > 	Please note that an association has two SCTP end points and these end
> > 	points have the IP addresses. So changing IP addresses means modifying
> > 	the end points.
> > 
> > 	Best regards
> > 	Michael  
> 
> -- 
> Brian F. G. Bidulock    > ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Apr 29 09:03:01 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA02345
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 09:03:01 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA19403
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 09:03:03 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA17869;
	Mon, 29 Apr 2002 08:37:42 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA17830
	for <tsvwg@optimus.ietf.org>; Mon, 29 Apr 2002 08:37:40 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-230.dallas.net [209.44.42.230])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA29542
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 08:37:36 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id HAA23117;
	Mon, 29 Apr 2002 07:37:15 -0500
Date: Mon, 29 Apr 2002 07:37:15 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
Cc: Jacob Heitz <jheitz@lucent.com>, "Khan, Mohammad" <mo_khan@trillium.com>,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
Message-ID: <20020429073715.E21316@openss7.org>
Reply-To: bidulock@openss7.org
References: <F92978223B01D311ACFF0008C71EE19D015DF7DC@MCHH225E>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <F92978223B01D311ACFF0008C71EE19D015DF7DC@MCHH225E>; from Michael.Tuexen@icn.siemens.de on Mon, Apr 29, 2002 at 02:20:03PM +0200
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Tuexen,

It is absurd, because it is an artificial restriction without merit
and without parallel in TCP.  Removing (ignoring) this sentence causes
no problems for the protocol.  Also, the peer has no way of detecting
or enforcing this restriction, making it even more absurd.

The other way of looking at it is to say that E1 and E1' *are* the
same "SCTP Endpoint", but are different communications endpoints (i.e.,
they are engaged in different associations).  Then even this absurd
statement is not violated (E1 and E1' are the same "SCTP Endpoint")
yet IP addresses can be added on a per-association basis.

--brian

On Mon, 29 Apr 2002, Tuexen Michael wrote:

> Brian,
> 
> see my comments below.
> 
> Best regards
> Michael
> 
> Michael Tuexen    Tel.:   +49 89 722 47210
> Siemens AG        Fax:    +49 89 722 48212
> ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de
> 
> > -----Original Message-----
> > From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> > Sent:	Monday, April 29, 2002 2:15 PM
> > To:	Tuexen Michael
> > Cc:	Jacob Heitz; Khan, Mohammad; 'tsvwg@ietf.org'; Pillai, Hari
> > Subject:	Re: [Tsvwg] EndPoint and Association Address conflict
> > 
> > Tuexen,
> > 
> > On Mon, 29 Apr 2002, Tuexen Michael wrote:
> > > > 
> > > > Why restrict Add-IP to an endpoint basis?
> > > > 
> > > 	[Tuexen Michael]  Consider E1 = ({IP'}, Port), E2 = ({IP'''}, Port),
> > > 	E3 = ({IP''''}, Port) with pairwise different IP', IP''', IP'''' and two associations
> > > 	A1: E1 <-> E2
> > > 	A2: E1 <-> E3
> > > 
> > > 	Now if we add IP'' to the association A1 we get
> > > 	A1 : E1' <-> E2
> > > 	with E1' = ({IP', IP''}, Port).
> > > 	E1' and E2 share one transport address which is in violation to RFC
> > > 	2960, page 13.
> > 
> > There is only one MUST on page 13: "Two SCTP endpoints MUST NOT have more
> > than one SCTP association between them at any given time."
> > 
> > How does this violate anything on page 13?
> 	[Tuexen Michael]  I did not mean this sentence. 
> > E1' and E2 only have one association between them.  E1 and E3 only have one
> > association between them.  Everthing appears ok as far as page 13 is
> > concerned.
> > 
> > Did you mean another page?
> > 
> > Or, did you mean the statement: "A transport address used by an
> > SCTP endpoint must not be used by another SCTP endpoint.  In other
> > words, a transport address is unique to an SCTP endpoint."  ...which
> > is absurd, particularly as it is done in TCP all the time, and when
> > you consider that non-multihomed SCTP associations need to reuse
> > dynamic port numbers for connection to other hosts.
> 	[Tuexen Michael]  I mean this sentence. This sentence
> 	is NOT absurd, at least if you use the definition of
> 	SCTP end-point I have given earlier. This is one of the reasons
> 	I'm quite sure that the RFC authors had the defintion I gave
> 	in mind.
> 
> > --brian
> > 
> > > 
> > > 	Please note that an association has two SCTP end points and these end
> > > 	points have the IP addresses. So changing IP addresses means modifying
> > > 	the end points.
> > > 
> > > 	Best regards
> > > 	Michael  
> > 
> > -- 
> > Brian F. G. Bidulock    > ¦ The reasonable man adapts himself to the ¦
> > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> >                         ¦ Therefore  all  progress  depends on the ¦
> >                         ¦ unreasonable man. -- George Bernard Shaw ¦

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Apr 29 09:58:57 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA08712
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 09:58:47 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA22546
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 09:58:49 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA20339;
	Mon, 29 Apr 2002 09:27:36 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA20255
	for <tsvwg@optimus.ietf.org>; Mon, 29 Apr 2002 09:27:31 -0400 (EDT)
Received: from gorilla.mchh.siemens.de (gorilla.mchh.siemens.de [194.138.158.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA05129
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 09:27:28 -0400 (EDT)
Received: from moody.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.226])
	by gorilla.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id PAA13255;
	Mon, 29 Apr 2002 15:27:20 +0200 (MET DST)
Received: from mchh247e.demchh201e.icn.siemens.de ([139.21.200.57])
	by moody.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id PAA21244;
	Mon, 29 Apr 2002 15:27:22 +0200 (MET DST)
Received: by MCHH247E with Internet Mail Service (5.5.2653.19)
	id <268H6VNA>; Mon, 29 Apr 2002 15:27:21 +0200
Message-ID: <F92978223B01D311ACFF0008C71EE19D015DF7DE@MCHH225E>
From: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
To: "'bidulock@openss7.org'" <bidulock@openss7.org>,
        Tuexen Michael
	 <Michael.Tuexen@icn.siemens.de>
Cc: Jacob Heitz <jheitz@lucent.com>, "Khan, Mohammad"
	 <mo_khan@trillium.com>,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: RE: [Tsvwg] EndPoint and Association Address conflict
Date: Mon, 29 Apr 2002 15:27:19 +0200
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Brian,
see my comments below.
Best regards
Michael

Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> Sent:	Monday, April 29, 2002 2:37 PM
> To:	Tuexen Michael
> Cc:	Jacob Heitz; Khan, Mohammad; 'tsvwg@ietf.org'; Pillai, Hari
> Subject:	Re: [Tsvwg] EndPoint and Association Address conflict
> 
> Tuexen,
> 
> It is absurd, because it is an artificial restriction without merit
> and without parallel in TCP.  Removing (ignoring) this sentence causes
> no problems for the protocol.  Also, the peer has no way of detecting
> or enforcing this restriction, making it even more absurd.
> 
	[Tuexen Michael]  It is just a technical term
	to talk about SCTP related stuff. There is no 
	'higher' meaning.
> The other way of looking at it is to say that E1 and E1' *are* the
> same "SCTP Endpoint", but are different communications endpoints (i.e.,
> they are engaged in different associations).  Then even this absurd
> statement is not violated (E1 and E1' are the same "SCTP Endpoint")
> yet IP addresses can be added on a per-association basis.
> 
	[Tuexen Michael]  E1 and E1' are not the same 
	end point. They have changes. One end point of association
	A1 has changed. AddIp, DelIP is about modification of
	end points.


> --brian
> 
> On Mon, 29 Apr 2002, Tuexen Michael wrote:
> 
> > Brian,
> > 
> > see my comments below.
> > 
> > Best regards
> > Michael
> > 
> > Michael Tuexen    Tel.:   +49 89 722 47210
> > Siemens AG        Fax:    +49 89 722 48212
> > ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de
> > 
> > > -----Original Message-----
> > > From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> > > Sent:	Monday, April 29, 2002 2:15 PM
> > > To:	Tuexen Michael
> > > Cc:	Jacob Heitz; Khan, Mohammad; 'tsvwg@ietf.org'; Pillai, Hari
> > > Subject:	Re: [Tsvwg] EndPoint and Association Address conflict
> > > 
> > > Tuexen,
> > > 
> > > On Mon, 29 Apr 2002, Tuexen Michael wrote:
> > > > > 
> > > > > Why restrict Add-IP to an endpoint basis?
> > > > > 
> > > > 	[Tuexen Michael]  Consider E1 = ({IP'}, Port), E2 = ({IP'''}, Port),
> > > > 	E3 = ({IP''''}, Port) with pairwise different IP', IP''', IP'''' and two associations
> > > > 	A1: E1 <-> E2
> > > > 	A2: E1 <-> E3
> > > > 
> > > > 	Now if we add IP'' to the association A1 we get
> > > > 	A1 : E1' <-> E2
> > > > 	with E1' = ({IP', IP''}, Port).
> > > > 	E1' and E2 share one transport address which is in violation to RFC
> > > > 	2960, page 13.
> > > 
> > > There is only one MUST on page 13: "Two SCTP endpoints MUST NOT have more
> > > than one SCTP association between them at any given time."
> > > 
> > > How does this violate anything on page 13?
> > 	[Tuexen Michael]  I did not mean this sentence. 
> > > E1' and E2 only have one association between them.  E1 and E3 only have one
> > > association between them.  Everthing appears ok as far as page 13 is
> > > concerned.
> > > 
> > > Did you mean another page?
> > > 
> > > Or, did you mean the statement: "A transport address used by an
> > > SCTP endpoint must not be used by another SCTP endpoint.  In other
> > > words, a transport address is unique to an SCTP endpoint."  ...which
> > > is absurd, particularly as it is done in TCP all the time, and when
> > > you consider that non-multihomed SCTP associations need to reuse
> > > dynamic port numbers for connection to other hosts.
> > 	[Tuexen Michael]  I mean this sentence. This sentence
> > 	is NOT absurd, at least if you use the definition of
> > 	SCTP end-point I have given earlier. This is one of the reasons
> > 	I'm quite sure that the RFC authors had the defintion I gave
> > 	in mind.
> > 
> > > --brian
> > > 
> > > > 
> > > > 	Please note that an association has two SCTP end points and these end> 
> > > > 	points have the IP addresses. So changing IP addresses means modifying
> > > > 	the end points.
> > > > 
> > > > 	Best regards
> > > > 	Michael  
> > > 
> > > -- 
> > > Brian F. G. Bidulock    > ¦ The reasonable man adapts himself to the ¦
> > > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> > >                         ¦ Therefore  all  progress  depends on the ¦
> > >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> 
> -- 
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Apr 29 10:02:10 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA09086
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 10:02:05 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA23117
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 10:02:07 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA20976;
	Mon, 29 Apr 2002 09:33:17 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA20947
	for <tsvwg@optimus.ietf.org>; Mon, 29 Apr 2002 09:33:15 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-230.dallas.net [209.44.42.230])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA05739
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 09:33:08 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id IAA24018;
	Mon, 29 Apr 2002 08:32:44 -0500
Date: Mon, 29 Apr 2002 08:32:44 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
Cc: Jacob Heitz <jheitz@lucent.com>, "Khan, Mohammad" <mo_khan@trillium.com>,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
Message-ID: <20020429083244.B15434@openss7.org>
Reply-To: bidulock@openss7.org
References: <F92978223B01D311ACFF0008C71EE19D015DF7DE@MCHH225E>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <F92978223B01D311ACFF0008C71EE19D015DF7DE@MCHH225E>; from Michael.Tuexen@icn.siemens.de on Mon, Apr 29, 2002 at 03:27:19PM +0200
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Michael,

Please see comments below...

On Mon, 29 Apr 2002, Tuexen Michael wrote:

> Brian,
> see my comments below.
> Best regards
> Michael
> 
> Michael Tuexen    Tel.:   +49 89 722 47210
> Siemens AG        Fax:    +49 89 722 48212
> ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de
> 
> > -----Original Message-----
> > From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> > Sent:	Monday, April 29, 2002 2:37 PM
> > To:	Tuexen Michael
> > Cc:	Jacob Heitz; Khan, Mohammad; 'tsvwg@ietf.org'; Pillai, Hari
> > Subject:	Re: [Tsvwg] EndPoint and Association Address conflict
> > 
> > Tuexen,
> > 
> > It is absurd, because it is an artificial restriction without merit
> > and without parallel in TCP.  Removing (ignoring) this sentence causes
> > no problems for the protocol.  Also, the peer has no way of detecting
> > or enforcing this restriction, making it even more absurd.
> > 
> 	[Tuexen Michael]  It is just a technical term
> 	to talk about SCTP related stuff. There is no 
> 	'higher' meaning.
> > The other way of looking at it is to say that E1 and E1' *are* the
> > same "SCTP Endpoint", but are different communications endpoints (i.e.,
> > they are engaged in different associations).  Then even this absurd
> > statement is not violated (E1 and E1' are the same "SCTP Endpoint")
> > yet IP addresses can be added on a per-association basis.
> > 
> 	[Tuexen Michael]  E1 and E1' are not the same 
> 	end point. They have changes. One end point of association
> 	A1 has changed. AddIp, DelIP is about modification of
> 	end points.

They are the same endpoint, because by definition a transport
address is unique to an endpoint.

--brian


> 
> 
> > --brian
> > 
> > On Mon, 29 Apr 2002, Tuexen Michael wrote:
> > 
> > > Brian,
> > > 
> > > see my comments below.
> > > 
> > > Best regards
> > > Michael
> > > 
> > > Michael Tuexen    Tel.:   +49 89 722 47210
> > > Siemens AG        Fax:    +49 89 722 48212
> > > ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de
> > > 
> > > > -----Original Message-----
> > > > From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> > > > Sent:	Monday, April 29, 2002 2:15 PM
> > > > To:	Tuexen Michael
> > > > Cc:	Jacob Heitz; Khan, Mohammad; 'tsvwg@ietf.org'; Pillai, Hari
> > > > Subject:	Re: [Tsvwg] EndPoint and Association Address conflict
> > > > 
> > > > Tuexen,
> > > > 
> > > > On Mon, 29 Apr 2002, Tuexen Michael wrote:
> > > > > > 
> > > > > > Why restrict Add-IP to an endpoint basis?
> > > > > > 
> > > > > 	[Tuexen Michael]  Consider E1 = ({IP'}, Port), E2 = ({IP'''}, Port),
> > > > > 	E3 = ({IP''''}, Port) with pairwise different IP', IP''', IP'''' and two associations
> > > > > 	A1: E1 <-> E2
> > > > > 	A2: E1 <-> E3
> > > > > 
> > > > > 	Now if we add IP'' to the association A1 we get
> > > > > 	A1 : E1' <-> E2
> > > > > 	with E1' = ({IP', IP''}, Port).
> > > > > 	E1' and E2 share one transport address which is in violation to RFC
> > > > > 	2960, page 13.
> > > > 
> > > > There is only one MUST on page 13: "Two SCTP endpoints MUST NOT have more
> > > > than one SCTP association between them at any given time."
> > > > 
> > > > How does this violate anything on page 13?
> > > 	[Tuexen Michael]  I did not mean this sentence. 
> > > > E1' and E2 only have one association between them.  E1 and E3 only have one
> > > > association between them.  Everthing appears ok as far as page 13 is
> > > > concerned.
> > > > 
> > > > Did you mean another page?
> > > > 
> > > > Or, did you mean the statement: "A transport address used by an
> > > > SCTP endpoint must not be used by another SCTP endpoint.  In other
> > > > words, a transport address is unique to an SCTP endpoint."  ...which
> > > > is absurd, particularly as it is done in TCP all the time, and when
> > > > you consider that non-multihomed SCTP associations need to reuse
> > > > dynamic port numbers for connection to other hosts.
> > > 	[Tuexen Michael]  I mean this sentence. This sentence
> > > 	is NOT absurd, at least if you use the definition of
> > > 	SCTP end-point I have given earlier. This is one of the reasons
> > > 	I'm quite sure that the RFC authors had the defintion I gave
> > > 	in mind.
> > > 
> > > > --brian
> > > > 
> > > > > 
> > > > > 	Please note that an association has two SCTP end points and these end> 
> > > > > 	points have the IP addresses. So changing IP addresses means modifying
> > > > > 	the end points.
> > > > > 
> > > > > 	Best regards
> > > > > 	Michael  
> > > > 
> > > > -- 
> > > > Brian F. G. Bidulock    > ¦ The reasonable man adapts himself to the ¦
> > > > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > > > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> > > >                         ¦ Therefore  all  progress  depends on the ¦
> > > >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> > 
> > -- 
> > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> >                         ¦ Therefore  all  progress  depends on the ¦
> >                         ¦ unreasonable man. -- George Bernard Shaw ¦

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Apr 29 10:02:28 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA09174
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 10:02:28 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA23195
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 10:02:27 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA21219;
	Mon, 29 Apr 2002 09:35:37 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA21189
	for <tsvwg@optimus.ietf.org>; Mon, 29 Apr 2002 09:35:33 -0400 (EDT)
Received: from beamer.mchh.siemens.de (beamer.mchh.siemens.de [194.138.158.163])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA06014
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 09:35:30 -0400 (EDT)
Received: from blues.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.227])
	by beamer.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id PAA00751;
	Mon, 29 Apr 2002 15:35:11 +0200 (MET DST)
Received: from mchh273e.demchh201e.icn.siemens.de ([139.21.200.83])
	by blues.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id PAA20071;
	Mon, 29 Apr 2002 15:34:45 +0200 (MET DST)
Received: by MCHH273E with Internet Mail Service (5.5.2653.19)
	id <JV3Q08HC>; Mon, 29 Apr 2002 15:35:20 +0200
Message-ID: <F92978223B01D311ACFF0008C71EE19D015DF7DF@MCHH225E>
From: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
To: "'bidulock@openss7.org'" <bidulock@openss7.org>,
        Tuexen Michael
	 <Michael.Tuexen@icn.siemens.de>
Cc: Jacob Heitz <jheitz@lucent.com>, "Khan, Mohammad"
	 <mo_khan@trillium.com>,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: RE: [Tsvwg] EndPoint and Association Address conflict
Date: Mon, 29 Apr 2002 15:35:15 +0200
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Brian,

see my comment below.

Best regards
Michael

Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> Sent:	Monday, April 29, 2002 3:33 PM
> To:	Tuexen Michael
> Cc:	Jacob Heitz; Khan, Mohammad; 'tsvwg@ietf.org'; Pillai, Hari
> Subject:	Re: [Tsvwg] EndPoint and Association Address conflict
> 
> Michael,
> 
> Please see comments below...
> 
> On Mon, 29 Apr 2002, Tuexen Michael wrote:
> 
> > Brian,
> > see my comments below.
> > Best regards
> > Michael
> > 
> > Michael Tuexen    Tel.:   +49 89 722 47210
> > Siemens AG        Fax:    +49 89 722 48212
> > ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de
> > 
> > > -----Original Message-----
> > > From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> > > Sent:	Monday, April 29, 2002 2:37 PM
> > > To:	Tuexen Michael
> > > Cc:	Jacob Heitz; Khan, Mohammad; 'tsvwg@ietf.org'; Pillai, Hari
> > > Subject:	Re: [Tsvwg] EndPoint and Association Address conflict
> > > 
> > > Tuexen,
> > > 
> > > It is absurd, because it is an artificial restriction without merit
> > > and without parallel in TCP.  Removing (ignoring) this sentence causes
> > > no problems for the protocol.  Also, the peer has no way of detecting
> > > or enforcing this restriction, making it even more absurd.
> > > 
> > 	[Tuexen Michael]  It is just a technical term
> > 	to talk about SCTP related stuff. There is no 
> > 	'higher' meaning.
> > > The other way of looking at it is to say that E1 and E1' *are* the
> > > same "SCTP Endpoint", but are different communications endpoints (i.e.,
> > > they are engaged in different associations).  Then even this absurd
> > > statement is not violated (E1 and E1' are the same "SCTP Endpoint")
> > > yet IP addresses can be added on a per-association basis.
> > > 
> > 	[Tuexen Michael]  E1 and E1' are not the same 
> > 	end point. They have changes. One end point of association
> > 	A1 has changed. AddIp, DelIP is about modification of
> > 	end points.
> 
> They are the same endpoint, because by definition a transport
> address is unique to an endpoint.
> 
	[Tuexen Michael]  No they are not the same, because they have not
	the same list of IP addresses. They have on in common. This is not a 
	violation to page 13 of RFC 2960, because they do not exist at the same
	time. When E1' exists, E1 does not exist anymore.
> --brian
> 
> 
> > 
> > 
> > > --brian
> > > 
> > > On Mon, 29 Apr 2002, Tuexen Michael wrote:
> > > 
> > > > Brian,
> > > > 
> > > > see my comments below.
> > > > 
> > > > Best regards
> > > > Michael
> > > > 
> > > > Michael Tuexen    Tel.:   +49 89 722 47210
> > > > Siemens AG        Fax:    +49 89 722 48212
> > > > ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de
> > > > 
> > > > > -----Original Message-----
> > > > > From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> > > > > Sent:	Monday, April 29, 2002 2:15 PM
> > > > > To:	Tuexen Michael
> > > > > Cc:	Jacob Heitz; Khan, Mohammad; 'tsvwg@ietf.org'; Pillai, Hari
> > > > > Subject:	Re: [Tsvwg] EndPoint and Association Address conflict
> > > > > 
> > > > > Tuexen,
> > > > > 
> > > > > On Mon, 29 Apr 2002, Tuexen Michael wrote:
> > > > > > > 
> > > > > > > Why restrict Add-IP to an endpoint basis?
> > > > > > > 
> > > > > > 	[Tuexen Michael]  Consider E1 = ({IP'}, Port), E2 = ({IP'''}, Port),
> > > > > > 	E3 = ({IP''''}, Port) with pairwise different IP', IP''', IP'''' and two associations
> > > > > > 	A1: E1 <-> E2
> > > > > > 	A2: E1 <-> E3
> > > > > > 
> > > > > > 	Now if we add IP'' to the association A1 we get
> > > > > > 	A1 : E1' <-> E2
> > > > > > 	with E1' = ({IP', IP''}, Port).
> > > > > > 	E1' and E2 share one transport address which is in violation to RFC> 
> > > > > > 	2960, page 13.
> > > > > 
> > > > > There is only one MUST on page 13: "Two SCTP endpoints MUST NOT have more
> > > > > than one SCTP association between them at any given time."
> > > > > 
> > > > > How does this violate anything on page 13?
> > > > 	[Tuexen Michael]  I did not mean this sentence. 
> > > > > E1' and E2 only have one association between them.  E1 and E3 only have one
> > > > > association between them.  Everthing appears ok as far as page 13 is
> > > > > concerned.
> > > > > 
> > > > > Did you mean another page?
> > > > > 
> > > > > Or, did you mean the statement: "A transport address used by an
> > > > > SCTP endpoint must not be used by another SCTP endpoint.  In other
> > > > > words, a transport address is unique to an SCTP endpoint."  ...which
> > > > > is absurd, particularly as it is done in TCP all the time, and when
> > > > > you consider that non-multihomed SCTP associations need to reuse
> > > > > dynamic port numbers for connection to other hosts.
> > > > 	[Tuexen Michael]  I mean this sentence. This sentence
> > > > 	is NOT absurd, at least if you use the definition of
> > > > 	SCTP end-point I have given earlier. This is one of the reasons
> > > > 	I'm quite sure that the RFC authors had the defintion I gave
> > > > 	in mind.
> > > > 
> > > > > --brian
> > > > > 
> > > > > > 
> > > > > > 	Please note that an association has two SCTP end points and these end> 
> > > > > > 	points have the IP addresses. So changing IP addresses means modifying
> > > > > > 	the end points.
> > > > > > 
> > > > > > 	Best regards
> > > > > > 	Michael  
> > > > > 
> > > > > -- 
> > > > > Brian F. G. Bidulock    > ¦ The reasonable man adapts himself to the ¦
> > > > > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > > > > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> > > > >                         ¦ Therefore  all  progress  depends on the ¦
> > > > >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> > > 
> > > -- 
> > > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> > >                         ¦ Therefore  all  progress  depends on the ¦
> > >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> 
> -- 
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Apr 29 10:50:55 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA14787
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 10:50:42 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA26153
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 10:50:44 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA23157;
	Mon, 29 Apr 2002 10:02:20 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA23129
	for <tsvwg@optimus.ietf.org>; Mon, 29 Apr 2002 10:02:17 -0400 (EDT)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA09131
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 10:02:13 -0400 (EDT)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir02nok.ntc.nokia.com (esvir02nokt.ntc.nokia.com [172.21.143.34])
	by mgw-x2.nokia.com (Switch-2.2.1/Switch-2.2.0) with ESMTP id g3TE2UF06014
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 17:02:31 +0300 (EET DST)
Received: from esebh001.NOE.Nokia.com (unverified) by esvir02nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5a8f176160ac158f22078@esvir02nok.ntc.nokia.com>;
 Mon, 29 Apr 2002 17:02:10 +0300
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh001.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Mon, 29 Apr 2002 17:02:10 +0300
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] EndPoint and Association Address conflict
Date: Mon, 29 Apr 2002 17:02:10 +0300
Message-ID: <58AEDE27D48F884285007ABF827FF26308EFB4@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] EndPoint and Association Address conflict
Thread-Index: AcHverQCfadm/OQYR8+9D/BDXqs82QABorNg
To: <bidulock@openss7.org>, <Michael.Tuexen@icn.siemens.de>,
        <qxie1@email.mot.com>, <randall@stewart.chicago.il.us>,
        <s_balani@trillium.com>, <Qiaobing.Xie@motorola.com>,
        <mo_khan@trillium.com>, <h_pillai@trillium.com>, <jheitz@lucent.com>,
        <tsvwg@ietf.org>
X-OriginalArrivalTime: 29 Apr 2002 14:02:10.0662 (UTC) FILETIME=[74C22060:01C1EF86]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id KAA23130
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi all!

	See my comments below...

> -----Original Message-----
> From: ext Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> Sent: 29. April 2002 15:37
> To: Tuexen Michael
> Cc: Jacob Heitz; Khan, Mohammad; 'tsvwg@ietf.org'; Pillai, Hari
> Subject: Re: [Tsvwg] EndPoint and Association Address conflict
> 
> 
> Tuexen,
> 
> It is absurd, because it is an artificial restriction without merit
> and without parallel in TCP.  Removing (ignoring) this sentence causes
> no problems for the protocol.  Also, the peer has no way of detecting
> or enforcing this restriction, making it even more absurd.

	I don't think it is absurd at all. You need to have a group of IP addresses that "go together". Otherwise, how could you know which addresses you should include in the INIT ACK when an INIT chunk hits one of those addresses/port? Or when the user tells you to establish a new association from endpoint X...

	And the restriction is completely logical... Otherwise, if you receive an INIT going to IPA/PortZ, and that address/port pair is part of two different endpoints, how can you know to which one it goes?

	As far as I understood from this long thread, Brian, you are saying that it is legal to have endpoint E1: [(IPA, IPB); Port 1024], and endpoint E2: [(IPA, IPC); Port 1024] (please, correct me if I'm wrong). Then, you receive an INIT (or any other chunk, but let's say it is an INIT to make things clearer), from endpoint EZ, going to IPA/Port 1024... Should you include IPA, IPB, or IPA, IPC?... That's why you cannot include the same address/port pair in two different endpoints...

	In my understanding, an endpoint is basically a port plus some of the host's addresses, and it is mainly used when responding to datagrams that don't go to any established association. It is kind of listener socket... When you receive something whose source/destination address/port is not part of any existing association, then you try to see if that datagram goes or not to any of the endpoint that might be "listening" to the destination address/port of the datagram (our source address/port). If there is any endpoint listening at the port to which the datagram was sent, then you process the datagram (as stated in section 8.4 of RFC 2960).

	And now, regarding the AddIP issues... Let me show you the way I see all this...

	In my understanding, when you establish a new association, you start with the "default" set of addresses, i.e., all of the addresses of the endpoint. Later on, you can delete some of those addresses in *some* associations, stopping using them as your source addresses. Also, if the endpoint adds new addresses, then, you can add those addresses to the associations that were created inside that endpoint...

	Basically, an association created from your endpoint E1, CAN NOT have more addresses than the ones that currently E1 has, but it CAN have less.

	In the examples given by Randall when you add a new card, I think that you have a problem if you bind to INADDR_ANY... That's because you don't know if the peer supports or not AddIP... If it doesn't, and you are not able to add the new addresses, you might end up receiving an ABORT from you peer if you use the newly added address.

	So I think that the associations should have knowledge about which addresses they can use as source address, because they might not always be all the addresses of the endpoint to which they belong... :-/

	And the point is not only that I might not be able to add the address, I might simply not want to do it. If it cost more money to send packets through the new address, then I might not want to use it at all, even though I could...

	So, associations can (independently) use AddIP extension to delete a source address in any case (except if it is the last one, of course). BUT, associations might only be able to add a new source address if the endpoint is using it. So, in my mind, I make a distinction between adding/deleting an address to/from an endpoint, and adding/deleting an address to/from a particular association... If you add an address to an association using ASCONFs, you MUST previously add it to the endpoint. If you delete an address from an endpoint, you SHOULD try to delete it previously using ASCONFs for all the associations using that source address and belonging to the endpoint to which the address is about to be added (i.e., also using the source address and the source port of the endpoint). I'm not sure if it should be a MUST or a SHOULD. If you are really unplugging the network card, then it doesn't matter since you are not really going to use that address any more either for sending or rec!
eiving. If you are simply doing it because you want to modify the endpoint (but the address is still physically working), then you should take care of what you are doing... :-/

	Summarizing:

	- An endpoint can use a subset of the source addresses of the host.
	- An association, can use a subset of the source addresses of the endpoint to which it belongs.

	I would very much appreciate comments about this mail...

	Thanks a lot!

	BR Iván Arias Rodríguez

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Apr 29 11:21:28 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA17752
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 11:21:27 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA28257
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 11:21:30 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA25547;
	Mon, 29 Apr 2002 10:38:52 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA25518
	for <tsvwg@optimus.ietf.org>; Mon, 29 Apr 2002 10:38:50 -0400 (EDT)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA13562
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 10:38:46 -0400 (EDT)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir04nok.ntc.nokia.com (esvir04nokt.ntc.nokia.com [172.21.143.36])
	by mgw-x2.nokia.com (Switch-2.2.1/Switch-2.2.0) with ESMTP id g3TEd7F25796
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 17:39:07 +0300 (EET DST)
Received: from esebh004.NOE.Nokia.com (unverified) by esvir04nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5a8f38e717ac158f24077@esvir04nok.ntc.nokia.com>;
 Mon, 29 Apr 2002 17:38:47 +0300
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh004.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Mon, 29 Apr 2002 17:38:47 +0300
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] T3 cannot timeout???
Date: Mon, 29 Apr 2002 17:38:47 +0300
Message-ID: <58AEDE27D48F884285007ABF827FF26308EFB5@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] T3 cannot timeout???
Thread-Index: AcHuLPgwFu/dzZDJTQ+t9/LPAfrCcQBWsK6A
To: <mohammed.asharaf@wipro.com>, <Dave.Sclarsky@radisys.com>
Cc: <tsvwg@ietf.org>, <Simon.Zhuang@radisys.com>, <Heinz.Prantner@radisys.com>,
        <Gene.Litt@radisys.com>, <Seamus.Gilchrist@radisys.com>
X-OriginalArrivalTime: 29 Apr 2002 14:38:47.0531 (UTC) FILETIME=[9231BBB0:01C1EF8B]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id KAA25519
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi Mohammed and the others!


> Shudn't the timer be started against the destination on which 
> this lowest
> TSN is retransmitted ( i.e here Dest :2  )?

	Now that I saw your comment, I think that the 4) apart of section 7.2.4 could be more clear than it is... In fact I'm not sure that I got it in the right way...

	In my opinion, the point here is that if you are fast retransmitting a TSN which was the first outstanding TSN on the address to which that TSN was originally sent, then you restart the T3 for that address to which the TSN was previously sent... Also you should restart it if the SACK acknowledged the first TSN outstanding in that address (this is the normal behavior, though).

	However, it could also been understood as that you restart the T3 timer of the address used for the fast retransmission, either if the first outstanding TSN in that address was acknowledged (i.e., the normal behavior), or if you are fast retransmitting a TSN to that address which is older than any of the TSNs that are already outstanding in that address... :-/

	Of course, a T3 timer should be running only if there are outstanding TSNs in that address... So if the first outstanding TSN is acknowledged (in any of the two options stated above), you only restart the timer if there are more TSNs outstanding in the address... otherwise you reset it...

	I think that the first option makes more sense (maybe because that's the one I understood :->). However, the second one makes also sense for me... :-/ The first option basically assumes that if the T3 (of the old address) started to run when that first TSN was sent, now that the TSN is not outstanding in that previous address any more, it makes sense to restart it... or even reset it, if the fast retransmitted TSN was the only one outstanding TSN in that address...

	However, option 2 says that, as when the T3 expires you start retransmitting the first outstanding TSNs, as in the "normal case" you will have DATA chunks of one MTU worth of data, normally you end up retransmitting only the first TSN... So it makes sense to "give extra time" when you send to an address a TSN which is the oldest outstanding TSN in that address... This also makes sense to me...

	Maybe there is a third way of understanding this, which is that you apply this to both the old and the new address... :-)

	What do people in the list think?

	In any case, as this section is being modified by the implementor's guide, it makes sense to state it clearer...

	Thanks!!

	BR Iván Arias Rodríguez

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 13:48:03 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA04912
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 13:48:02 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA07372
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 13:48:04 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA05392;
	Mon, 29 Apr 2002 13:07:42 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA05346
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 13:07:38 -0400 (EDT)
Received: from hermes.fm.intel.com (fmr01.intel.com [192.55.52.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA29220
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 13:07:34 -0400 (EDT)
Received: from talaria.fm.intel.com (talaria.fm.intel.com [10.1.192.39])
	by hermes.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.42 2002/04/26 23:25:13 root Exp $) with ESMTP id g3TH8W016004
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 17:08:32 GMT
Received: from fmsmsxvs042.fm.intel.com (fmsmsxv042-1.fm.intel.com [132.233.48.110])
	by talaria.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.17 2002/04/27 00:24:04 root Exp $) with SMTP id g3TH9sI22110
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 17:09:54 GMT
Received: from FMSMSX018.fm.intel.com ([132.233.42.197])
 by fmsmsxvs042.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002042910115201331
 ; Mon, 29 Apr 2002 10:11:52 -0700
Received: by fmsmsx018.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <2WWTWV7S>; Mon, 29 Apr 2002 10:07:34 -0700
Message-ID: <F1CE15E08172D4119247009027AE9D50191173AB@FMSMSX37>
From: "Balani, Sandeep" <s_balani@trillium.com>
To: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        Qiaobing Xie <qxie1@email.mot.com>
Cc: "Balani, Sandeep" <s_balani@trillium.com>,
        "'Qiaobing Xie'"<Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Mon, 29 Apr 2002 10:07:32 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

hi Randy, 
  Please See my comments ....

> -----Original Message-----
> From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: Saturday, April 27, 2002 5:59 AM
> To: Qiaobing Xie
> Cc: Balani, Sandeep; 'Qiaobing Xie'; Ivan.Arias-Rodriguez@nokia.com;
> Khan, Mohammad; tsvwg@ietf.org
> Subject: Re: [Tsvwg] End point and association Address conflict
> 
> 
> Qiabing/All:
> 
> I agree 100% with Q here. We need to tighten the wording up
> but I don't think anything needs to be redefined... Now I
> want to bring up something for all of you to consider... 
> 
> FORGET ADD-IP for a MINUTE...
> 
> 
> In a NORMAL RFC2960 association I have the following
> 
> E1-(Bound INADDR_ANY)
> 
> I currently have
> 
> IP-A: Nic1/IP-B: Nic2
> 
> I setup an association with a peer
> 
> A1=(IP-A,IP-B:Port 4444)<====================>(IP-Z:Port 5555)
> 
> Now everything is fine...
> 
> Now My administrator plugs in Nic3 and does:
> 
> ifconfig nic3 IP-C
> 
> Now ... my endpoint has JUST CHANGED!! It is now
> 
> IP-A: Nic1/IP-B: Nic2/IP-C: Nic3
> 
> (remember we are bound to ALL)....
> 
> Now  my peer does 
> 
>                  <<------TO:IP-C Port4444----INIT(IP:Z port 5555)
> 
> .....
> 
> You have the SAME EXACT thing you have been dicussing ..
> 
> 
> This is one of the reasons we NEED Add-IP. 
> 
> ....If we think about the ADD-IP case... then when the ADDRESS
> was added to the ENDPOINT (automatically) ... you would have gotten
> 
> -----ADD-IP(IP-C)-------->
> 
> and not reached the above condition...
> 
[ Sandeep ] I agree, that ADD-IP in this case has saved us. But Just a small
question, What happens that if this endpoint E1 has associations A1 & A2,
And A1 supports ADD-IP & A2 Does *NOT* support ADD-IP, what should be done
for A2 after IPc nic was added ? 

I think disconnect A2 and re-establish it again, any other suggestion or
option ??????? 

If this is the case Network administrator should know the consequences(
disconnecting some associations )of ADDING IP addresses dynamically, Is'nt
it ? 

> The key is has Qiaobing stated, you ADD addresses to ENDPOINTs. I
> think we will add some text towards this clearfication in the next
> pass..
> 
 [ Sandeep] I am glad that this is being cleared UP, So the same statment is
TRUE for DELETE-IP also, I.E.  " You DELETE addresses from an Endpoint,
Although final beneficiary in both the cases is Association, but
ADDITION/DELETION of IP addresses is done on association."  Correct
????????????????????  


If everything said above is TRUE than what i am gonna assume is  " When User
Does ADDIP i will add that address in all the assocation for that Endpoint,
& When User does DELETE IP i will delete the address from all the
association "  

However i have real problem if some association does'nt support ADD/DELETE
IP procedures then what :-(   

> But the BIG thing I want to point our here is that this issue is
> NOT because of ADD-IP but is something that can happen WITHOUT
> ADD-IP... This is VERY important to remember... Add-ip actually
> helps this situation... as long as your implementation of it
> adds addresses to endpoints that is :->
> 
> R
> 
> Qiaobing Xie wrote:
> > 
> > Hi, Sandeep,
> > 
> > > [ Sandeep ] I totally totally agree with U on that " 
> Addition/Deletion of IP
> > > Address should be on an Endpoint & not on Association.  
> After all we are
> > > modifying the source IP addresses which is an essential 
> part of Endpoint "
> > >
> > > But Currently ADDIP/DELETEIP Draft Says :
> > >
> > > "
> > > - Dynamic IP address reconfiguration extension: For
> > >     computational or networking platforms that allow 
> addition/removal of
> > >     physical interface cards this feature can provide:
> > >
> > >      A) a graceful method to add to the interfaces of an existing
> > >         *******association*******. For IPv6 this feature 
> allows renumbering
> > >         of existing associations.
> > > "
> > 
> > I agree with you. These are loose wordings which may need to be
> > tightened up in the draft. But in a general sense, with the correct
> > understanding and clarification of the rules, the above 
> wording is still
> > fine - it simply depicts the final beneficiary of this add/delete-ip
> > feature, which is indeed the associations themselves.
> > 
> > More importantly, in the draft we need to explicit notify the
> > implementor that using add-ip/delete-ip is fine and 
> benefiting, but one
> > must not break the association rules, otherwise, it becomes 
> a misuse of
> > a good tool and you implementation will break.
> > 
> > I am glad this is brought up in this discussion thread.
> > 
> > >
> > > There are similar references like this in the DRAFT, 
> which mentions
> > > ADD/DELETE IP procedures assocaitions-wise.
> > 
> > We will improve the text and make it tighter.
> > 
> > > So Is there any plans to change the defination of 
> ADDIP/DELETEIP because If
> > > i would do ADD/DELETE on Endpoint I have to perform 
> ASCONF/ASCONF-ACK
> > > procedures on all the assocations  belonging to that 
> endpoint which seems
> > > correct to me but In that case I won't be  in compliance 
> with draft  ....
> > 
> > I don't think anything needs to be re-defined. We just need more
> > clarifications. To the implementor, a good add/delete-ip 
> API semantics
> > might help to provide a safeguard against misuses.
> > 
> > SCTP is different from TCP in that it allows a single endpoint to
> > associate with multiple peers (the "fan-out" config). But 
> no matter how
> > many peers the endpoint is concurrently connected to, 
> logically it is
> > _still_ a single endpoint.
> > 
> > With this being cleared, it should not be hard to 
> understand that it is
> > conceptually wrong to let a user apply add/delete-ip to 
> only a subset of
> > the fan-out concurrent connections on that single endpoint. 
> Because this
> > would in fact result in "splitting" the single endpoint into two
> > different ones. This is NEVER the intention of add/delete-ip - the
> > feature is meant to dynamically modify an endpoint without 
> tearing down
> > all the associations it currently has; it is NOT meant to 
> split a single
> > endpoint into two!
> > 
> > -Qiaobing
> > 
> > >
> > > > With this important clarification of concept, all above 
> should become
> > > > very simple and clear - what you are allowed to do here 
> is __to delete
> > > > IPb from endpoint E1 which is currently being used in 
> associations A1
> > > > and A9__. If your implementation is doing the right 
> thing, the result
> > > > should be
> > > >
> > > >                E1'             E2
> > > >     A1' = [(IPa; Port1), (IPz; Port 65535)]
> > > >
> > > >     And
> > > >                E1'                   E2  /* call it E3 */
> > > >     A9' = [ (IPa; Port1), (IPm; Port 76577)]
> > > >
> > > > If your intention is to transform the original A1 and A9 to the
> > > > following:
> > > >
> > > >                E1'             E2
> > > >     A1' = [(IPa; Port1), (IPz; Port 65535)]
> > > >
> > > >     And
> > > >                E1'                   E2  /* call it E3 */
> > > >     A9' = [ (IPb; Port1), (IPm; Port 76577)]
> > > >
> > > > I am afraid that you have to tear down A1/A9 and set-up 
> A1' and A9'
> > > > fresh. ADD/DELET-IP will NOT get your here without 
> violating the basic
> > > > association rules.
> > > >
> > > > -Qiaobing
> > > >
> > >
> > > Thanks
> > > Sandeep
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 14:22:13 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA09126
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 14:22:13 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA10462
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 14:22:15 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA07715;
	Mon, 29 Apr 2002 13:56:41 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA07686
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 13:56:38 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA06354
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 13:56:35 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3THtCH59894;
	Mon, 29 Apr 2002 12:55:13 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCD8900.4547552A@stewart.chicago.il.us>
Date: Mon, 29 Apr 2002 12:55:12 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: bidulock@openss7.org, Michael.Tuexen@icn.siemens.de, qxie1@email.mot.com,
        s_balani@trillium.com, Qiaobing.Xie@motorola.com, mo_khan@trillium.com,
        h_pillai@trillium.com, jheitz@lucent.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFB4@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Ivan:

My only comment is ..

Yes, I agree with what you are saying and think you have
stated it quite well :>

R

Ivan.Arias-Rodriguez@nokia.com wrote:
> 
>         Hi all!
> 
>         See my comments below...
> 
> > -----Original Message-----
> > From: ext Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> > Sent: 29. April 2002 15:37
> > To: Tuexen Michael
> > Cc: Jacob Heitz; Khan, Mohammad; 'tsvwg@ietf.org'; Pillai, Hari
> > Subject: Re: [Tsvwg] EndPoint and Association Address conflict
> >
> >
> > Tuexen,
> >
> > It is absurd, because it is an artificial restriction without merit
> > and without parallel in TCP.  Removing (ignoring) this sentence causes
> > no problems for the protocol.  Also, the peer has no way of detecting
> > or enforcing this restriction, making it even more absurd.
> 
>         I don't think it is absurd at all. You need to have a group of IP addresses that "go together". Otherwise, how could you know which addresses you should include in the INIT ACK when an INIT chunk hits one of those addresses/port? Or when the user tells you to establish a new association from endpoint X...
> 
>         And the restriction is completely logical... Otherwise, if you receive an INIT going to IPA/PortZ, and that address/port pair is part of two different endpoints, how can you know to which one it goes?
> 
>         As far as I understood from this long thread, Brian, you are saying that it is legal to have endpoint E1: [(IPA, IPB); Port 1024], and endpoint E2: [(IPA, IPC); Port 1024] (please, correct me if I'm wrong). Then, you receive an INIT (or any other chunk, but let's say it is an INIT to make things clearer), from endpoint EZ, going to IPA/Port 1024... Should you include IPA, IPB, or IPA, IPC?... That's why you cannot include the same address/port pair in two different endpoints...
> 
>         In my understanding, an endpoint is basically a port plus some of the host's addresses, and it is mainly used when responding to datagrams that don't go to any established association. It is kind of listener socket... When you receive something whose source/destination address/port is not part of any existing association, then you try to see if that datagram goes or not to any of the endpoint that might be "listening" to the destination address/port of the datagram (our source address/port). If there is any endpoint listening at the port to which the datagram was sent, then you process the datagram (as stated in section 8.4 of RFC 2960).
> 
>         And now, regarding the AddIP issues... Let me show you the way I see all this...
> 
>         In my understanding, when you establish a new association, you start with the "default" set of addresses, i.e., all of the addresses of the endpoint. Later on, you can delete some of those addresses in *some* associations, stopping using them as your source addresses. Also, if the endpoint adds new addresses, then, you can add those addresses to the associations that were created inside that endpoint...
> 
>         Basically, an association created from your endpoint E1, CAN NOT have more addresses than the ones that currently E1 has, but it CAN have less.
> 
>         In the examples given by Randall when you add a new card, I think that you have a problem if you bind to INADDR_ANY... That's because you don't know if the peer supports or not AddIP... If it doesn't, and you are not able to add the new addresses, you might end up receiving an ABORT from you peer if you use the newly added address.
> 
>         So I think that the associations should have knowledge about which addresses they can use as source address, because they might not always be all the addresses of the endpoint to which they belong... :-/
> 
>         And the point is not only that I might not be able to add the address, I might simply not want to do it. If it cost more money to send packets through the new address, then I might not want to use it at all, even though I could...
> 
>         So, associations can (independently) use AddIP extension to delete a source address in any case (except if it is the last one, of course). BUT, associations might only be able to add a new source address if the endpoint is using it. So, in my mind, I make a distinction between adding/deleting an address to/from an endpoint, and adding/deleting an address to/from a particular association... If you add an address to an association using ASCONFs, you MUST previously add it to the endpoint. If you delete an address from an endpoint, you SHOULD try to delete it previously using ASCONFs for all the associations using that source address and belonging to the endpoint to which the address is about to be added (i.e., also using the source address and the source port of the endpoint). I'm not sure if it should be a MUST or a SHOULD. If you are really unplugging the network card, then it doesn't matter since you are not really going to use that address any more either for sen!
 ding or
> rec!
> eiving. If you are simply doing it because you want to modify the endpoint (but the address is still physically working), then you should take care of what you are doing... :-/
> 
>         Summarizing:
> 
>         - An endpoint can use a subset of the source addresses of the host.
>         - An association, can use a subset of the source addresses of the endpoint to which it belongs.
> 
>         I would very much appreciate comments about this mail...
> 
>         Thanks a lot!
> 
>         BR Iván Arias Rodríguez
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 14:24:52 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA09533
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 14:24:52 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA10620
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 14:24:51 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA08048;
	Mon, 29 Apr 2002 13:59:49 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA16644
	for <tsvwg@optimus.ietf.org>; Mon, 29 Apr 2002 08:24:40 -0400 (EDT)
Received: from gorilla.mchh.siemens.de (gorilla.mchh.siemens.de [194.138.158.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA28343
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 08:24:37 -0400 (EDT)
Received: from blues.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.227])
	by gorilla.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id OAA07823;
	Mon, 29 Apr 2002 14:24:37 +0200 (MET DST)
Received: from mchh273e.demchh201e.icn.siemens.de ([139.21.200.83])
	by blues.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id OAA03672;
	Mon, 29 Apr 2002 14:24:01 +0200 (MET DST)
Received: by MCHH273E with Internet Mail Service (5.5.2653.19)
	id <JV3Q0Z5D>; Mon, 29 Apr 2002 14:24:36 +0200
Message-ID: <F92978223B01D311ACFF0008C71EE19D015DF7DD@MCHH225E>
From: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
To: "'bidulock@openss7.org'" <bidulock@openss7.org>,
        Tuexen Michael
	 <Michael.Tuexen@icn.siemens.de>
Cc: Qiaobing Xie <qxie1@email.mot.com>,
        "Balani, Sandeep"
	 <s_balani@trillium.com>,
        "'Qiaobing Xie'" <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, randall@stewart.chicago.il.us,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Mon, 29 Apr 2002 14:24:30 +0200
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Brian,

looking at the RFC I say that RFC 2960 uses the
definition I have posted before. But that is my
understanding, you can have a different one. 

I think you are implying to much from the term
end point used in natural language. Look at it as
a technical term 'SCTP end point' as only defined in RFC
2960.

This only indicates that there might be some clarification
necessary in the impl. guide.

Best regards
Michael

Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> Sent:	Monday, April 29, 2002 2:16 PM
> To:	Tuexen Michael
> Cc:	Qiaobing Xie; Balani, Sandeep; 'Qiaobing Xie'; Ivan.Arias-Rodriguez@nokia.com; randall@stewart.chicago.il.us; Khan, Mohammad; tsvwg@ietf.org
> Subject:	Re: [Tsvwg] End point and association Address conflict
> 
> Michael,
> 
> You may as well let the RFC talk for itself, then.
> 
> --brian
> 
> On Mon, 29 Apr 2002, Tuexen Michael wrote:
> 
> > Brian,
> > 
> > I'm talking ONLY about an 'SCTP endpoint', as being defined in
> > RFC 2960. I agree, that the term endpoint is used with a lot
> > of meanings in natural language. I'm using it not implying any
> > other meaning (for example for TCP or so), it is just a term
> > defined in RFC 2960. 
> > 
> > I do understand your examples and I only say that the notion 
> > defined in RFC 2960 does not apply to them in a way you are using
> > them. 
> > 
> > Maybe you think that is the wrong notion, but I think that is the 
> > one used in RFC 2960 and the extension IDs. Since it is necessary
> > to use the same one, Randy or Qiaobing could state which one they
> > had in mind when writing the RFC. Then we can all use the same.
> > 
> > Best regards
> > Michael
> > Michael Tuexen    Tel.:   +49 89 722 47210
> > Siemens AG        Fax:    +49 89 722 48212
> > ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de
> > 
> > > -----Original Message-----
> > > From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> > > Sent:	Monday, April 29, 2002 1:31 PM
> > > To:	Tuexen Michael
> > > Cc:	Qiaobing Xie; Balani, Sandeep; 'Qiaobing Xie'; Ivan.Arias-Rodriguez@nokia.com; randall@stewart.chicago.il.us; Khan, Mohammad; tsvwg@ietf.org
> > > Subject:	Re: [Tsvwg] End point and association Address conflict
> > > 
> > > Michael,
> > > 
> > > Please see comments below...
> > > 
> > > On Mon, 29 Apr 2002, Tuexen Michael wrote:
> > > 
> > > > > On Mon, 29 Apr 2002, Tuexen Michael wrote:
> > > > > 
> > > > > > Hi Brian,
> > > > > > 
> > > > > > if a host A has two IP-addresses IP' and IP'' and two associations
> > > > > > from (IP', Port) to a peer (IP, Port) and (IP'', Port) to the peer
> > > > > > (IP, Port) then 
> > > > > > - host A has two IP-addresses
> > > > > > - there are two end-points (IP', Port), (IP'', Port), each of them
> > > > > >   uses only one IP address available at host A.
> > > > > > This 'splits' (in your terminology) host A into two end points.
> > > > > > The difference between IP addresses of the end point and IP addresses
> > > > > > of the host is also the reason why it is valid to say
> > > > > > - you MUST include all IP address of the end point in the INIT INIT-ACK.
> > > > > > - you include a subset of all addresses of the host in the INIT INIT-ACK.
> > > > > 
> > > > > Why do you call (IPA, Port) one endpoint?  I view an endpoint as
> > > > > half an association.  When two associations are formed from (IPA, Port),
> > > > > there are two endpoints, not one.  When adding IPB and IPC and can
> > > > > add IPB to one endpoint to form (IPB, IPA, Port) and IPC to another
> > > > > to form (IPC, IPA, Port).  Why restrict this?  Particularly if that is
> > > > > the way I would have bound them if IPB and IPC where available in the
> > > > > first place?
> > > > 
> > > > [Tuexen Michael]  I think we are not using the same definition of end> 
> > > > point. I'm using an abstract one based on RFC 2960. 
> > > > Def.: An SCTP end point is a pair of a set of IP addresses and a port
> > > >       number.
> > > 
> > > We all agree that an endpoint is a set of IP addresses and a port.  This
> > > does not say that an endpoint is a unique set of IP addresses and a
> > > port.  The cardinality of endpoints is not defined.
> > > 
> > > > This means, considering a server, I have only one end point for the
> > > > server.
> > > 
> > > Your interpretation.  I have given an example where more than one server
> > > can service the same endpoint.  This is typical and true even for TCP
> > > (where multiple HTTP servers, for example, service the same IP/Port).
> > > 
> > > > Mulitple clients can conncect to the server.  These associations share
> > > > the same end point.
> > > 
> > > Depending on how one views the cardinality of endpoints, these
> > > associations can be viewed as connecting to different endpoints (even
> > > though those endpoints were orginally bound to the same IP...IP/Port,
> > > that is they had the same name).
> > > 
> > > > 
> > > > This definition is in tune with the sentence "Two SCTP end points MUST
> > > > not have more than one SCTP association between them at any time
> > > > given" RFC 2960, page 13, defintion of SCTP association.
> > > 
> > > This gives no clues as to the cardinality of endpoints.  It only states
> > > the cardinality of associations.
> > > 
> > > > 
> > > > Using your definition as an end point is "half of an association" The
> > > > sentence cited abover would be trivial.
> > > 
> > > The sentence above is what makes an endpoint half an association...
> > > 
> > > > 
> > > > That is why I'm still thinking that my notion of SCTP end point is the
> > > > one the authors of RFC 2960 had in mind.
> > > 
> > > Endpoint has had far more before RFC 2960.  An endpoint for
> > > communication is the full duplex communications endpoint which occurs at
> > > either end of the association.  When we bind and endpoint (give it a
> > > name) we associate the IP...IP/Port with the endpoint.  When we connect
> > > an association, we form a unique connection to another endpoint.  That
> > > endpoint also has an name: IP...IP/Port, but we only contact it using
> > > one IP address.
> > > 
> > > Just as with TCP, if we make an active connection from an endpoint only
> > > (that is, we are not listening) then multiple sockets/streams can be
> > > bound (named) with the same IP...IP/Port combination.  In fact, with
> > > dynamic port allocation, IP/Port combinations are reused on TCP every
> > > day between multiple connecting TCP endpoints (sockets/streams) that
> > > have no clue that the magic port number `0' has yeilded the same IP/Port
> > > combination as another TCP endpoint (socket/stream).
> > > 
> > > This is also the case of SCTP.  I have two sockets which I have bound
> > > to IPA/IPB/Port`0'.  I connect from one socket to Host B and the
> > > implementation assigns me port 10000.  I connect from the other socket
> > > to Host C and the implementation also assigns me port 10000  (knowing
> > > that there is no conflict because one IPA/IPB/Port10000 socket is
> > > connected to HostB and the other IPA/IPB/Port10000 socket is connected
> > > to HostC).  Now how can you call these the same endpoint?
> > > 
> > > --brian
> > > 
> > > > 
> > > > Best regards
> > > > Michael
> > > > 
> > > 
> > > -- 
> > > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> > >                         ¦ Therefore  all  progress  depends on the ¦
> > >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> 
> -- 
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    > ¦>  worl> d; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 14:26:07 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA09721
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 14:26:02 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA10671
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 14:26:04 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA08384;
	Mon, 29 Apr 2002 14:00:49 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA04260
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 12:56:29 -0400 (EDT)
Received: from motgate2.mot.com (motgate2.mot.com [136.182.1.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA28393
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 12:56:27 -0400 (EDT)
Received: [from pobox4.mot.com (pobox4.mot.com [10.64.251.243]) by motgate2.mot.com (motgate2 2.1) with ESMTP id JAA07713; Mon, 29 Apr 2002 09:56:27 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox4.mot.com (MOT-pobox4 2.0) with ESMTP id JAA19036; Mon, 29 Apr 2002 09:56:26 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3TGuNj25186;
	Mon, 29 Apr 2002 11:56:23 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id LAA21662;
	Mon, 29 Apr 2002 11:56:33 -0500 (CDT)
Message-ID: <3CCD7BEA.67A009BD@email.mot.com>
Date: Mon, 29 Apr 2002 11:59:22 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
CC: bidulock@openss7.org, "Balani, Sandeep" <s_balani@trillium.com>,
        "'Qiaobing Xie'" <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, randall@stewart.chicago.il.us,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F92978223B01D311ACFF0008C71EE19D015DF7D9@MCHH225E>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Tuexen Michael wrote:
...snip...
> Maybe you think that is the wrong notion, but I think that is the
> one used in RFC 2960 and the extension IDs. Since it is necessary
> to use the same one, Randy or Qiaobing could state which one they
> had in mind when writing the RFC. Then we can all use the same.

I 100% agree with Michael and it is exactly why we have an endpoint
definition in RFC2960 - it is meant to tell the reader that when RFC2960
is concerned THIS IS THE DEFINITION TO USE. Any other "endpoint"
definition from other literature, no matter how elegant or ugly it might
be, is entirely IRRELEVANT :-)

-Qiaobing

> 
> Best regards
> Michael
> Michael Tuexen    Tel.:   +49 89 722 47210
> Siemens AG        Fax:    +49 89 722 48212
> ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de
>


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 14:27:27 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA09935
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 14:27:27 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA10735
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 14:27:29 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA07900;
	Mon, 29 Apr 2002 13:59:20 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA11651
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 06:53:41 -0400 (EDT)
Received: from gorilla.mchh.siemens.de (gorilla.mchh.siemens.de [194.138.158.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA21445
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 06:53:39 -0400 (EDT)
Received: from blues.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.227])
	by gorilla.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id MAA29448;
	Mon, 29 Apr 2002 12:53:36 +0200 (MET DST)
Received: from mchh274e.demchh201e.icn.siemens.de ([139.21.200.84])
	by blues.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id MAA13787;
	Mon, 29 Apr 2002 12:53:01 +0200 (MET DST)
Received: by MCHH274E with Internet Mail Service (5.5.2653.19)
	id <268QLAGV>; Mon, 29 Apr 2002 12:53:35 +0200
Message-ID: <F92978223B01D311ACFF0008C71EE19D015DF7D7@MCHH225E>
From: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
To: "'bidulock@openss7.org'" <bidulock@openss7.org>,
        Tuexen Michael
	 <Michael.Tuexen@icn.siemens.de>
Cc: Qiaobing Xie <qxie1@email.mot.com>,
        "Balani, Sandeep"
	 <s_balani@trillium.com>,
        "'Qiaobing Xie'" <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, randall@stewart.chicago.il.us,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Mon, 29 Apr 2002 12:53:33 +0200
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Brian,

see my comments below.

Best regards
Michael

Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> Sent:	Monday, April 29, 2002 11:01 AM
> To:	Tuexen Michael
> Cc:	Qiaobing Xie; Balani, Sandeep; 'Qiaobing Xie'; Ivan.Arias-Rodriguez@nokia.com; randall@stewart.chicago.il.us; Khan, Mohammad; tsvwg@ietf.org
> Subject:	Re: [Tsvwg] End point and association Address conflict
> 
> Michael,
> 
> On Mon, 29 Apr 2002, Tuexen Michael wrote:
> 
> > Hi Brian,
> > 
> > if a host A has two IP-addresses IP' and IP'' and two associations
> > from (IP', Port) to a peer (IP, Port) and (IP'', Port) to the peer
> > (IP, Port) then 
> > - host A has two IP-addresses
> > - there are two end-points (IP', Port), (IP'', Port), each of them
> >   uses only one IP address available at host A.
> > This 'splits' (in your terminology) host A into two end points.
> > The difference between IP addresses of the end point and IP addresses
> > of the host is also the reason why it is valid to say
> > - you MUST include all IP address of the end point in the INIT INIT-ACK.
> > - you include a subset of all addresses of the host in the INIT INIT-ACK.
> 
> Why do you call (IPA, Port) one endpoint?  I view an endpoint as
> half an association.  When two associations are formed from (IPA, Port),
> there are two endpoints, not one.  When adding IPB and IPC and can
> add IPB to one endpoint to form (IPB, IPA, Port) and IPC to another
> to form (IPC, IPA, Port).  Why restrict this?  Particularly if that is
> the way I would have bound them if IPB and IPC where available in the
> first place?
	[Tuexen Michael]  I think we are not using the same
	definition of end point. I'm using an abstract one based
	on RFC 2960. 
	Def.: An SCTP end point is a pair of a set of IP
	      addresses and a port number.
	This means, considering a server, I have only one end point
	for the server. Mulitple clients can conncect to the server.
	These associations share the same end point.

	This definition is in tune with the sentence
	"Two SCTP end points MUST not have more than one
	 SCTP association between them at any time given"
	RFC 2960, page 13, defintion of SCTP association.

	Using your definition as an end point is "half of an 
	association" The sentence cited abover would be trivial.

	That is why I'm still thinking that my notion of SCTP
	end point is the one the authors of RFC 2960 had in mind.

	Best regards
	Michael


> I have a host that talks to two other hosts.  IPA is a default and
> backup interface on this host.  IPB and IPC are dedicated high-speed
> interfaces for talking to HostB and HostC on an engineered network.
> Normally I would bind applications serving HostB to IPB as primary
> and IPA as secondary.  I would bind applications serving HostC to IPC
> as primary and IPA as secondary.  I have a default server which binds
> to IPA as primary.  All three servers listen on Port X.
> 
> With IPB and IPC down, the default A-server is listening on IPA Port X.
> When HostB connects, it spawns a B-server and forms the association
> (IPA, Port X) to HostB.  When HostC connects, it spawns an C-server
> and forms the association (IPA, Port X) to HostC.  When IPB becomes
> availble, the B-server wants it added to its association.  The C-server
> does not.  When IPC becomes available, the C-server wants it added to
> its association.  The B-Server does not.
> 
> So, unless the listener or connecter is actually bound to INADDR_ANY,
> there is no reason to force it to add IP addresses just because some
> other association happens to be engaged in an association with the
> same local addresses and port numbers.> 
> 
> --brian
> 
> 
> > 
> > Best regards
> > Michael
> > 
> > Michael Tuexen    Tel.:   +49 89 722 47210
> > Siemens AG        Fax:    +49 89 722 48212
> > ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de
> > 
> > > -----Original Message-----
> > > From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> > > Sent:	Monday, April 29, 2002 8:50 AM
> > > To:	Qiaobing Xie
> > > Cc:	Balani, Sandeep; 'Qiaobing Xie'; Ivan.Arias-Rodriguez@nokia.com; randall@stewart.chicago.il.us; Khan, Mohammad; tsvwg@ietf.org
> > > Subject:	Re: [Tsvwg] End point and association Address conflict
> > > 
> > > Qiaobing,
> > > 
> > > Multi-homing is OPTIONAL.  Are you saying that a group of
> > > associations that are not multi-homed (single IP address
> > > in INIT), and on different IP addresses for the same host
> > > is somehow splitting the host into multiple end-points?
> > > I don't see how this is different than for TCP when the
> > > host has multiple IP addresses: each connection may be
> > > from a different IP address...  Please explain.
> > > 
> > > --brian
> > > 
> > > On Fri, 26 Apr 2002, Qiaobing Xie wrote:
> > > 
> > > > 
> > > > I don't think anything needs to be re-defined. We just need more
> > > > clarifications. To the implementor, a good add/delete-ip API semantics
> > > > might help to provide a safeguard against misuses. 
> > > > 
> > > > SCTP is different from TCP in that it allows a single endpoint to
> > > > associate with multiple peers (the "fan-out" config). But no matter how
> > > > many peers the endpoint is concurrently connected to, logically it is
> > > > _still_ a single endpoint. 
> > > > 
> > > > With this being cleared, it should not be hard to understand that it is
> > > > conceptually wrong to let a user apply add/delete-ip to only a subset of
> > > > the fan-out concurrent connections on that single endpoint. Because this
> > > > would in fact result in "splitting" the single endpoint into two
> > > > different ones. This is NEVER the intention of add/delete-ip - the
> > > > feature is meant to dynamically modify an endpoint without tearing down
> > > > all the associations it currently has; it is NOT meant to split a single
> > > > endpoint into two!
> > > > 
> > > > -Qiaobing
> > > > 
> > > 
> > > -- 
> > > Brian F. G. Bidulock    > ¦ The reasonable man adapts himself to the ¦
> > > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> > >                         ¦ Therefore  all  progress  depends on the ¦
> > >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> > > 
> > > _______________________________________________
> > > tsvwg mailing list
> > > tsvwg@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> -- 
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 14:30:48 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA10297
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 14:30:48 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA11167
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 14:30:50 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA08581;
	Mon, 29 Apr 2002 14:01:15 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA07350
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 13:47:49 -0400 (EDT)
Received: from motgate.mot.com (motgate.mot.com [129.188.136.100])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA04838
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 13:47:46 -0400 (EDT)
Received: [from pobox3.mot.com (pobox3.mot.com [10.64.251.242]) by motgate.mot.com (motgate 2.1) with ESMTP id KAA00836 for <tsvwg@ietf.org>; Mon, 29 Apr 2002 10:47:48 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox3.mot.com (MOT-pobox3 2.0) with ESMTP id KAA21135 for <tsvwg@ietf.org>; Mon, 29 Apr 2002 10:35:05 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3THlij00427;
	Mon, 29 Apr 2002 12:47:44 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id MAA21715;
	Mon, 29 Apr 2002 12:47:54 -0500 (CDT)
Message-ID: <3CCD87F2.245B633E@email.mot.com>
Date: Mon, 29 Apr 2002 12:50:42 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: "Khan, Mohammad" <mo_khan@trillium.com>
CC: tsvwg@ietf.org, "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <A1171C2850D8D4119D7B0002A508C764016983C9@bgsmsx91.iind.intel.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi, kashif,

I didn't see your this email before I sent out my "rules" email a few
min ago. I totally agree with your following rules. 

-Qiaobing

"Khan, Mohammad" wrote:
> 
> Hi All,
> I could make the following observations from the ongoing discussions about
> Endpoint and Association address conflict for SCTP.
> 
> 1)INIT chunk shall contain all the addresses bound to the Endpoint.
> 
> 2)Addition and deletion of IP address shall be on Endpoint not on
> Association.
> 
> For the point 2) suppose an Endpoint E1(IPA and Port X) has multiple
> associations A1, A2 and A3 with three different peers.
> 
>         E1(IPA and PortX)                       Peer Hosts
>         A1(E1)                                  IPB and Port X
>         A2(E1)                                  IPC and Port X
>         A3(E1)                                  IPD and Port X
> As per the assumption 2) to add an IPZ at E1 SCTP has to send ASCONF chunk
> on all the A1, A2 and A3 associations and it shall not consider this address
> as added till it receives success from all the associations. If failure is
> received is received from any of the associations then this Add-IP request
> shall be discarded.
> 
> Same procedure is to be adopted to delete an address from an endpoint.
> Please comment.
> 
> Regards
> kashif
> 
> 
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 14:33:22 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA10380
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 14:33:22 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA11643
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 14:33:20 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA07942;
	Mon, 29 Apr 2002 13:59:32 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA13898
	for <tsvwg@optimus.ietf.org>; Mon, 29 Apr 2002 07:41:39 -0400 (EDT)
Received: from gorilla.mchh.siemens.de (gorilla.mchh.siemens.de [194.138.158.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA23821
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 07:41:37 -0400 (EDT)
Received: from blues.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.227])
	by gorilla.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id NAA03791;
	Mon, 29 Apr 2002 13:41:36 +0200 (MET DST)
Received: from mchh273e.demchh201e.icn.siemens.de ([139.21.200.83])
	by blues.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id NAA24427;
	Mon, 29 Apr 2002 13:41:01 +0200 (MET DST)
Received: by MCHH273E with Internet Mail Service (5.5.2653.19)
	id <JV3Q0X61>; Mon, 29 Apr 2002 13:41:36 +0200
Message-ID: <F92978223B01D311ACFF0008C71EE19D015DF7D9@MCHH225E>
From: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
To: "'bidulock@openss7.org'" <bidulock@openss7.org>,
        Tuexen Michael
	 <Michael.Tuexen@icn.siemens.de>
Cc: Qiaobing Xie <qxie1@email.mot.com>,
        "Balani, Sandeep"
	 <s_balani@trillium.com>,
        "'Qiaobing Xie'" <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, randall@stewart.chicago.il.us,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Mon, 29 Apr 2002 13:41:31 +0200
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Brian,

I'm talking ONLY about an 'SCTP endpoint', as being defined in
RFC 2960. I agree, that the term endpoint is used with a lot
of meanings in natural language. I'm using it not implying any
other meaning (for example for TCP or so), it is just a term
defined in RFC 2960. 

I do understand your examples and I only say that the notion 
defined in RFC 2960 does not apply to them in a way you are using
them. 

Maybe you think that is the wrong notion, but I think that is the 
one used in RFC 2960 and the extension IDs. Since it is necessary
to use the same one, Randy or Qiaobing could state which one they
had in mind when writing the RFC. Then we can all use the same.

Best regards
Michael
Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> Sent:	Monday, April 29, 2002 1:31 PM
> To:	Tuexen Michael
> Cc:	Qiaobing Xie; Balani, Sandeep; 'Qiaobing Xie'; Ivan.Arias-Rodriguez@nokia.com; randall@stewart.chicago.il.us; Khan, Mohammad; tsvwg@ietf.org
> Subject:	Re: [Tsvwg] End point and association Address conflict
> 
> Michael,
> 
> Please see comments below...
> 
> On Mon, 29 Apr 2002, Tuexen Michael wrote:
> 
> > > On Mon, 29 Apr 2002, Tuexen Michael wrote:
> > > 
> > > > Hi Brian,
> > > > 
> > > > if a host A has two IP-addresses IP' and IP'' and two associations
> > > > from (IP', Port) to a peer (IP, Port) and (IP'', Port) to the peer
> > > > (IP, Port) then 
> > > > - host A has two IP-addresses
> > > > - there are two end-points (IP', Port), (IP'', Port), each of them
> > > >   uses only one IP address available at host A.
> > > > This 'splits' (in your terminology) host A into two end points.
> > > > The difference between IP addresses of the end point and IP addresses
> > > > of the host is also the reason why it is valid to say
> > > > - you MUST include all IP address of the end point in the INIT INIT-ACK.
> > > > - you include a subset of all addresses of the host in the INIT INIT-ACK.
> > > 
> > > Why do you call (IPA, Port) one endpoint?  I view an endpoint as
> > > half an association.  When two associations are formed from (IPA, Port),
> > > there are two endpoints, not one.  When adding IPB and IPC and can
> > > add IPB to one endpoint to form (IPB, IPA, Port) and IPC to another
> > > to form (IPC, IPA, Port).  Why restrict this?  Particularly if that is
> > > the way I would have bound them if IPB and IPC where available in the
> > > first place?
> > 
> > [Tuexen Michael]  I think we are not using the same definition of end
> > point. I'm using an abstract one based on RFC 2960. 
> > Def.: An SCTP end point is a pair of a set of IP addresses and a port
> >       number.
> 
> We all agree that an endpoint is a set of IP addresses and a port.  This
> does not say that an endpoint is a unique set of IP addresses and a
> port.  The cardinality of endpoints is not defined.
> 
> > This means, considering a server, I have only one end point for the
> > server.
> 
> Your interpretation.  I have given an example where more than one server
> can service the same endpoint.  This is typical and true even for TCP
> (where multiple HTTP servers, for example, service the same IP/Port).
> 
> > Mulitple clients can conncect to the server.  These associations share
> > the same end point.
> 
> Depending on how one views the cardinality of endpoints, these
> associations can be viewed as connecting to different endpoints (even
> though those endpoints were orginally bound to the same IP...IP/Port,
> that is they had the same name).
> 
> > 
> > This definition is in tune with the sentence "Two SCTP end points MUST
> > not have more than one SCTP association between them at any time
> > given" RFC 2960, page 13, defintion of SCTP association.
> 
> This gives no clues as to the cardinality of endpoints.  It only states
> the cardinality of associations.
> 
> > 
> > Using your definition as an end point is "half of an association" The
> > sentence cited abover would be trivial.
> 
> The sentence above is what makes an endpoint half an association...
> 
> > 
> > That is why I'm still thinking that my notion of SCTP end point is the
> > one the authors of RFC 2960 had in mind.
> 
> Endpoint has had far more before RFC 2960.  An endpoint for
> communication is the full duplex communications endpoint which occurs at
> either end of the association.  When we bind and endpoint (give it a
> name) we associate the IP...IP/Port with the endpoint.  When we connect
> an association, we form a unique connection to another endpoint.  That
> endpoint also has an name: IP...IP/Port, but we only contact it using
> one IP address.
> 
> Just as with TCP, if we make an active connection from an endpoint only
> (that is, we are not listening) then multiple sockets/streams can be
> bound (named) with the same IP...IP/Port combination.  In fact, with
> dynamic port allocation, IP/Port combinations are reused on TCP every
> day between multiple connecting TCP endpoints (sockets/streams) that
> have no clue that the magic port number `0' has yeilded the same IP/Port
> combination as another TCP endpoint (socket/stream).
> 
> This is also the case of SCTP.  I have two sockets which I have bound
> to IPA/IPB/Port`0'.  I connect from one socket to Host B and the
> implementation assigns me port 10000.  I connect from the other socket
> to Host C and the implementation also assigns me port 10000  (knowing
> that there is no conflict because one IPA/IPB/Port10000 socket is
> connected to HostB and the other IPA/IPB/Port10000 socket is connected
> to HostC).  Now how can you call these the same endpoint?
> 
> --brian
> 
> > 
> > Best regards
> > Michael
> > 
> 
> -- 
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 14:33:35 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA10402
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 14:33:35 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA11690
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 14:33:37 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA08180;
	Mon, 29 Apr 2002 14:00:15 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA03693
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 12:41:19 -0400 (EDT)
Received: from motgate.mot.com (motgate.mot.com [129.188.136.100])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA26397
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 12:41:17 -0400 (EDT)
Received: [from pobox3.mot.com (pobox3.mot.com [10.64.251.242]) by motgate.mot.com (motgate 2.1) with ESMTP id JAA03070; Mon, 29 Apr 2002 09:41:17 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox3.mot.com (MOT-pobox3 2.0) with ESMTP id JAA20041; Mon, 29 Apr 2002 09:28:30 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3TGf5j23840;
	Mon, 29 Apr 2002 11:41:05 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id LAA21642;
	Mon, 29 Apr 2002 11:41:15 -0500 (CDT)
Message-ID: <3CCD7853.7E0941B8@email.mot.com>
Date: Mon, 29 Apr 2002 11:44:03 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: bidulock@openss7.org
CC: Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        "Balani, Sandeep" <s_balani@trillium.com>,
        Ivan.Arias-Rodriguez@nokia.com, randall@stewart.chicago.il.us,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F1CE15E08172D4119247009027AE9D50191173A9@FMSMSX37> <3CC9E579.7DB84355@email.mot.com> <20020429014931.A17185@openss7.org>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Hi, Brian,

The entire chapter 2 of Randy and I's SCTP book is spent on defining
SCTP endpoints. If you have it handy, take a look at Figure 2-2 on page
23 and Figure 2-3 on page 24 (too hard for me to re-draw them in text
here). That shows what "split" endpoints look like.

-Qiaobing

"Brian F. G. Bidulock" wrote:
> 
> Qiaobing,
> 
> Multi-homing is OPTIONAL.  Are you saying that a group of
> associations that are not multi-homed (single IP address
> in INIT), and on different IP addresses for the same host
> is somehow splitting the host into multiple end-points?
> I don't see how this is different than for TCP when the
> host has multiple IP addresses: each connection may be
> from a different IP address...  Please explain.
> 
> --brian
> 
> On Fri, 26 Apr 2002, Qiaobing Xie wrote:
> 
> >
> > I don't think anything needs to be re-defined. We just need more
> > clarifications. To the implementor, a good add/delete-ip API semantics
> > might help to provide a safeguard against misuses.
> >
> > SCTP is different from TCP in that it allows a single endpoint to
> > associate with multiple peers (the "fan-out" config). But no matter how
> > many peers the endpoint is concurrently connected to, logically it is
> > _still_ a single endpoint.
> >
> > With this being cleared, it should not be hard to understand that it is
> > conceptually wrong to let a user apply add/delete-ip to only a subset of
> > the fan-out concurrent connections on that single endpoint. Because this
> > would in fact result in "splitting" the single endpoint into two
> > different ones. This is NEVER the intention of add/delete-ip - the
> > feature is meant to dynamically modify an endpoint without tearing down
> > all the associations it currently has; it is NOT meant to split a single
> > endpoint into two!
> >
> > -Qiaobing
> >
> 
> --
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 14:34:26 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA10458
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 14:34:26 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA11775
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 14:34:28 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA08455;
	Mon, 29 Apr 2002 14:01:00 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA07018
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 13:37:21 -0400 (EDT)
Received: from motgate.mot.com (motgate.mot.com [129.188.136.100])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA02128
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 13:37:19 -0400 (EDT)
Received: [from pobox2.mot.com (pobox2.mot.com [136.182.15.8]) by motgate.mot.com (motgate 2.1) with ESMTP id KAA22429; Mon, 29 Apr 2002 10:37:20 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox2.mot.com (MOT-pobox2 2.0) with ESMTP id KAA18921; Mon, 29 Apr 2002 10:37:20 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3THbCj29556;
	Mon, 29 Apr 2002 12:37:13 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id MAA21697;
	Mon, 29 Apr 2002 12:37:23 -0500 (CDT)
Message-ID: <3CCD857A.7902D0D@email.mot.com>
Date: Mon, 29 Apr 2002 12:40:11 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
CC: "'Qiaobing Xie'" <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, randall@stewart.chicago.il.us,
        s_balani@trillium.com, mo_khan@trillium.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F92978223B01D311ACFF0008C71EE19D015DF7D2@MCHH225E>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi, Michael,

I think two simple rules will be sufficient. Let me use an example to
illustrate them.

Assume we have the following endpoints and associations in place:

on local host         on remote host(s)
A1: E1 <-----------------> E3 
A2: E1 <---------------------> E4
A3: E2 <-------------------------> E5

Rule 1: whenever adding an SCTP transport to a local endpoint, make sure
this same transport is NOT already present in other existing endpoints
(e.g., you don't want to add one of E1's transport to E2).

Rule 2: when adding a transport to a local endpoint that has
associations to multiple peers (ie, a fan-out config), one MUST send an
add-ip ASCONF to each of those peers. And the new transport can only be
used until ALL the peers ack successfully. If anyone of the peers denies
the add-ip then you can't add it. So, if you add a new transport to E1
above, you have to send ASCONF to both E3 and E4 and have to get a
positive ack from both of them.

-Qiaobing

Tuexen Michael wrote:
> 
> Dear all,
> 
> looking at the discussion it seems that we have only one restriction
> for add-IP:
> 
> Consider the set E of all SCTP end points at a host. Partition
> E = E_1 union E_2 union ... E_65535
> into disjoint sets E_P of all end points having port P.
> 
> When adding an IP-address to multiple end points, all of these
> end points MUST belong to different E_P.
> 
> This means that an SCTP implementation has to choose from all end points
> having the same port on end point to which the add IP procedure is applied.
> 
> Another point:
> Consider two associations
> A1: E1 <-> E2  (The SCTP implementation of E2 is AddIP capable.)
> A2: E1 <-> E3  (The SCTP implementation of E3 is not AddIP capable.)
> 
> Now you add an IP address to E1 and get E1'. So A1 is now E1' <-> E2.
> Then we consider A2 also as A2 : E1' <-> E3, but the stack will never
> use the added address. Is this correct usage of terminology?
> 
> Best regards
> Michael
> 
> Michael Tuexen    Tel.:   +49 89 722 47210
> Siemens AG        Fax:    +49 89 722 48212
> ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de
> 
> > -----Original Message-----
> > From: Qiaobing Xie [SMTP:qxie1@email.mot.com]
> > Sent: Friday, April 26, 2002 2:45 AM
> > To:   Ivan.Arias-Rodriguez@nokia.com
> > Cc:   randall@stewart.chicago.il.us; s_balani@trillium.com; Qiaobing Xie; mo_khan@trillium.com; tsvwg@ietf.org
> > Subject:      Re: [Tsvwg] End point and association Address conflict
> >
> > Ivan.Arias-Rodriguez@nokia.com wrote:
> > ....
> > >         Now I have another doubt... Imagine that we have an endpoint
> > > E1 with addresses IPa, IPb and port 1. We have an association A1
> > > that only uses IPa, whose peer is IPz and port 65535. Thus the
> > > association A1 can be represented as [(IPa; Port1), (IPz; Port 65535)].
> > > What happens now if we receive an INIT from IPz/65535 directed to IPb/1?
> > >
> > >         As IPb is not being used in A1, we should answer with an
> > > INIT ACK as in a normal establishment of an association. However,
> > > when you receive an INIT chunk, you include all the addresses of the endpoint
> >                                       ^^^^^^^^^^^^^
> > Here your "..include all the addr..." directly violated the "under no
> > circumstance can an SCTP transport address ever be present in more than
> > one SCTP endpoint" rule (see last the paragraph, page 23 of the SCTP
> > book).
> >
> >         E1               E2
> > A1 = [(IPa; Port1), (IPz; Port 65535)]
> >
> >         E3                 E2
> > A2 = [(IPa, IPb; Port1), (IPz; Port 65535)]
> >
> > And you have IPa/Port1 present in both E1 and E3 (they are TWO different
> > endpoints by definition!). This is prohibited by the SCTP association
> > definition.
> >
> > -Qiaobing
> >
> > > in the INIT ACK chunk... But we have to be careful, because in
> > > this case, when receiving the COOKIE ECHO, we have two options:
> >
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 14:36:24 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA10578
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 14:36:24 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA11944
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 14:36:26 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA08806;
	Mon, 29 Apr 2002 14:01:55 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA07846
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 13:59:07 -0400 (EDT)
Received: from motgate4.mot.com (motgate4.mot.com [144.189.100.102])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA06587
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 13:59:04 -0400 (EDT)
Received: [from pobox.mot.com (pobox.mot.com [129.188.137.100]) by motgate4.mot.com (motgate4 2.1) with ESMTP id KAA28751; Mon, 29 Apr 2002 10:59:05 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox.mot.com (MOT-pobox 2.0) with ESMTP id KAA07594; Mon, 29 Apr 2002 10:59:04 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3THwwj01334;
	Mon, 29 Apr 2002 12:58:58 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id MAA21729;
	Mon, 29 Apr 2002 12:59:08 -0500 (CDT)
Message-ID: <3CCD8A95.72572C78@email.mot.com>
Date: Mon, 29 Apr 2002 13:01:57 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: bidulock@openss7.org, Michael.Tuexen@icn.siemens.de,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        randall@stewart.chicago.il.us, s_balani@trillium.com,
        mo_khan@trillium.com, h_pillai@trillium.com, jheitz@lucent.com,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFB4@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Hi, Ivan:

Ivan.Arias-Rodriguez@nokia.com wrote:
....snip....
> 
>         Summarizing:
> 
>         - An endpoint can use a subset of the source addresses of the host.

Yes.

>         - An association, can use a subset of the source addresses of the 
>           endpoint to which it belongs.

No, this is incorrect. 

All associations coming out of an endpoint MUST use all the source
addresses of the endpoint all the time. However, the address set of the
endpoint can change over time (by adding or deleting IPs). Whenever that
happens ALL the existing associations coming out of the endpoint MUST be
updated to the new address set (this is why we have ASCONF exchanges).
Moreover, there are rules over how the address set of the endpoint can
be changed. That is what I sent out in my previous email.

regards,
-Qiaobing

> 
>         I would very much appreciate comments about this mail...
> 
>         Thanks a lot!
> 
>         BR Iván Arias Rodríguez


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 14:38:30 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA10623
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 14:38:30 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA12057
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 14:38:32 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA08264;
	Mon, 29 Apr 2002 14:00:30 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA03870
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 12:46:26 -0400 (EDT)
Received: from motgate2.mot.com (motgate2.mot.com [136.182.1.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA27103
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 12:46:24 -0400 (EDT)
Received: [from pobox.mot.com (pobox.mot.com [129.188.137.100]) by motgate2.mot.com (motgate2 2.1) with ESMTP id JAA29848 for <tsvwg@ietf.org>; Mon, 29 Apr 2002 09:46:25 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox.mot.com (MOT-pobox 2.0) with ESMTP id JAA04963 for <tsvwg@ietf.org>; Mon, 29 Apr 2002 09:46:25 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3TGkLj24343;
	Mon, 29 Apr 2002 11:46:21 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id LAA21651;
	Mon, 29 Apr 2002 11:46:31 -0500 (CDT)
Message-ID: <3CCD798F.35A7F77F@email.mot.com>
Date: Mon, 29 Apr 2002 11:49:19 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: bidulock@openss7.org
CC: Tuexen Michael <Michael.Tuexen@icn.siemens.de>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        "Balani, Sandeep" <s_balani@trillium.com>,
        Ivan.Arias-Rodriguez@nokia.com, randall@stewart.chicago.il.us,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F92978223B01D311ACFF0008C71EE19D015DF7D1@MCHH225E> <20020429040040.B19082@openss7.org>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

"Brian F. G. Bidulock" wrote:
...
> Why do you call (IPA, Port) one endpoint?  I view an endpoint as
> half an association.
...snip...

Brian, 

SCTP endpoints exist before an association. Your view of "an endpoint as
half an association" is incorrect.

-Qiaobing


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 14:41:16 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA10673
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 14:41:16 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA12187
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 14:41:18 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA09622;
	Mon, 29 Apr 2002 14:07:17 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA09596
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 14:07:15 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA07453
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 14:07:12 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3TI74H59975;
	Mon, 29 Apr 2002 13:07:05 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCD8BC8.CB90FA61@stewart.chicago.il.us>
Date: Mon, 29 Apr 2002 13:07:04 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: "Khan, Mohammad" <mo_khan@trillium.com>
CC: "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <A1171C2850D8D4119D7B0002A508C764016983C9@bgsmsx91.iind.intel.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

"Khan, Mohammad" wrote:
> As per the assumption 2) to add an IPZ at E1 SCTP has to send ASCONF chunk
> on all the A1, A2 and A3 associations and it shall not consider this address
> as added till it receives success from all the associations. If failure is
> received is received from any of the associations then this Add-IP request
> shall be discarded.
> 

Here I disagree.. You add it to the endpoint and it is added
RIGHT away..

At the same time you start sending off your ASCONF's. If for some reason
an association  fails to add an address (either via non-support of
add-ip OR some other failure) that association is in a odd state
until either:

A) it restarts

or

B) it terminates.

This is such a corner case I don't see it as an issue. The corner
case being that your peer (who does NOT support Add-IP) sends an
INIT to the new address. If that were to occur, when you unpacked
the cookie you would find the other association and in my opinion
to handle the corner case.. abort the newly forming association
(not the old one) or just drop the cookie and let it timeout until
it goes away...

I don't see this to be an issue...

R
-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 14:44:27 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA10813
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 14:44:27 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA12390
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 14:44:29 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA08685;
	Mon, 29 Apr 2002 14:01:30 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA07800
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 13:58:03 -0400 (EDT)
Received: from mg03.austin.ibm.com (mg03.austin.ibm.com [192.35.232.20])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA06482
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 13:58:00 -0400 (EDT)
Received: from austin.ibm.com (netmail1.austin.ibm.com [9.3.7.138])
	by mg03.austin.ibm.com (AIX4.3/8.9.3/8.9.3) with ESMTP id MAA31170;
	Mon, 29 Apr 2002 12:53:34 -0500
Received: from austin.ibm.com (sctptest.dynamic.austin.ibm.com [9.111.216.144] (may be forged))
	by austin.ibm.com (AIX4.3/8.9.3/8.9.3) with ESMTP id MAA38466;
	Mon, 29 Apr 2002 12:57:52 -0500
Message-ID: <3CCD844A.DB1BC960@austin.ibm.com>
Date: Mon, 29 Apr 2002 12:35:06 -0500
From: Jon Grimm <jgrimm@austin.ibm.com>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.4.18lksctp i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: bidulock@openss7.org, Michael.Tuexen@icn.siemens.de, qxie1@email.mot.com,
        randall@stewart.chicago.il.us, s_balani@trillium.com,
        Qiaobing.Xie@motorola.com, mo_khan@trillium.com, h_pillai@trillium.com,
        jheitz@lucent.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFB4@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Ivan/All
	
	I think something that needs to come out of this discussion are
scenarios and mandatory behavior in those scenarios (or why flexibility
is allowed if not mandatory behavior).  Its a little disturbing that a
similar 'what is an endpoint?' discussion seems to pop up again every so
often.  (e.g.bindx discussion last Sept in
ftp://ftp.ietf.org/ietf-mail-archive/tsvwg/2001-09.mail ) and still does
not seem to be completely clear even to those following SCTP closely
over the last year.  


For each of the following scenarios,

1) A new address is to be added an endpoint. (manual) 
2) An address is to be deleted from an endpoint.  (manual)
3) A new NIC/address is configured and an endpoint is bound to all
(auto)
4) An address bound is brought down (no longer available) (auto).   
5) An address bound is brought down for maintenance and then brought
on-line again. (auto)

consider each of the following subcases (assuming multiple associations
under an ep).  

a) All Peers have ADD-IP.  
b) No Peers have ADD-IP.
c) A subset of Peers support ADD-IP.


What happens?  


Another comment:
Dropping down a little bit into the Sockets API for illustration:

Depending on one's interpretation of endpoint, what is the endpoint for
the association of a peeled-off UDP style socket or for that matter an
'accepted' TCP-style socket.  If the strict interpretation of the
endpoint is kept.. wouldn't an ADD-IP on the child sockets (peeled off
or accept TCP) affect the local endpoint of the parent socket?  Or vise
versa, should an ADD-IP on a TCP-style listening socket, affect the
accepted sockets?  If yes, was this expected?  If no, does the original
definition of endpoint still hold?  

Even if each expected behavior does not get spelled it, it would be nice
for the SCTP Sockets API I-D to clarify the relationship between a
socket and an endpoint, so implementors can at least infer what the
intended behavior should be.   


Best Regards,
Jon Grimm

Ivan.Arias-Rodriguez@nokia.com wrote:

>         I don't think it is absurd at all. You need to have a group of IP addresses that "go together". Otherwise, how could you know which addresses you should include in the INIT ACK when an INIT chunk hits one of those addresses/port? Or when the user tells you to establish a new association from endpoint X...
> 
>         And the restriction is completely logical... Otherwise, if you receive an INIT going to IPA/PortZ, and that address/port pair is part of two different endpoints, how can you know to which one it goes?
> 
>         As far as I understood from this long thread, Brian, you are saying that it is legal to have endpoint E1: [(IPA, IPB); Port 1024], and endpoint E2: [(IPA, IPC); Port 1024] (please, correct me if I'm wrong). Then, you receive an INIT (or any other chunk, but let's say it is an INIT to make things clearer), from endpoint EZ, going to IPA/Port 1024... Should you include IPA, IPB, or IPA, IPC?... That's why you cannot include the same address/port pair in two different endpoints...
> 
>         In my understanding, an endpoint is basically a port plus some of the host's addresses, and it is mainly used when responding to datagrams that don't go to any established association. It is kind of listener socket... When you receive something whose source/destination address/port is not part of any existing association, then you try to see if that datagram goes or not to any of the endpoint that might be "listening" to the destination address/port of the datagram (our source address/port). If there is any endpoint listening at the port to which the datagram was sent, then you process the datagram (as stated in section 8.4 of RFC 2960).
> 
>         And now, regarding the AddIP issues... Let me show you the way I see all this...
> 
>         In my understanding, when you establish a new association, you start with the "default" set of addresses, i.e., all of the addresses of the endpoint. Later on, you can delete some of those addresses in *some* associations, stopping using them as your source addresses. Also, if the endpoint adds new addresses, then, you can add those addresses to the associations that were created inside that endpoint...
> 
>         Basically, an association created from your endpoint E1, CAN NOT have more addresses than the ones that currently E1 has, but it CAN have less.
> 
>         In the examples given by Randall when you add a new card, I think that you have a problem if you bind to INADDR_ANY... That's because you don't know if the peer supports or not AddIP... If it doesn't, and you are not able to add the new addresses, you might end up receiving an ABORT from you peer if you use the newly added address.
> 
>         So I think that the associations should have knowledge about which addresses they can use as source address, because they might not always be all the addresses of the endpoint to which they belong... :-/
> 
>         And the point is not only that I might not be able to add the address, I might simply not want to do it. If it cost more money to send packets through the new address, then I might not want to use it at all, even though I could...
> 
>         So, associations can (independently) use AddIP extension to delete a source address in any case (except if it is the last one, of course). BUT, associations might only be able to add a new source address if the endpoint is using it. So, in my mind, I make a distinction between adding/deleting an address to/from an endpoint, and adding/deleting an address to/from a particular association... If you add an address to an association using ASCONFs, you MUST previously add it to the endpoint. If you delete an address from an endpoint, you SHOULD try to delete it previously using ASCONFs for all the associations using that source address and belonging to the endpoint to which the address is about to be added (i.e., also using the source address and the source port of the endpoint). I'm not sure if it should be a MUST or a SHOULD. If you are really unplugging the network card, then it doesn't matter since you are not really going to use that address any more either for sen!
 di!
!
ng or receiving. If you are simply doing it because you want to modify the endpoint (but the address is still physically working), then you should take care of what you are doing... :-/
> 
>         Summarizing:
> 
>         - An endpoint can use a subset of the source addresses of the host.
>         - An association, can use a subset of the source addresses of the endpoint to which it belongs.
> 
>         I would very much appreciate comments about this mail...
> 
>         Thanks a lot!
> 
>         BR Iván Arias Rodríguez
>


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 14:47:42 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA11052
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 14:47:37 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA12521
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 14:47:39 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA08588;
	Mon, 29 Apr 2002 14:01:16 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA08516
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 14:01:10 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA06743
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 14:01:07 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3TI0dH59916;
	Mon, 29 Apr 2002 13:00:40 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCD8A47.96AB3987@stewart.chicago.il.us>
Date: Mon, 29 Apr 2002 13:00:39 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: "Balani, Sandeep" <s_balani@trillium.com>
CC: Qiaobing Xie <qxie1@email.mot.com>,
        "'Qiaobing Xie'" <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F1CE15E08172D4119247009027AE9D50191173AB@FMSMSX37>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

"Balani, Sandeep" wrote:
> 
> hi Randy,
>   Please See my comments ....
> 
> > -----Original Message-----
> > From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> > Sent: Saturday, April 27, 2002 5:59 AM
> > To: Qiaobing Xie
> > Cc: Balani, Sandeep; 'Qiaobing Xie'; Ivan.Arias-Rodriguez@nokia.com;
> > Khan, Mohammad; tsvwg@ietf.org
> > Subject: Re: [Tsvwg] End point and association Address conflict
> >
> >
> > Qiabing/All:
> >
> > I agree 100% with Q here. We need to tighten the wording up
> > but I don't think anything needs to be redefined... Now I
> > want to bring up something for all of you to consider...
> >
> > FORGET ADD-IP for a MINUTE...
> >
> >
> > In a NORMAL RFC2960 association I have the following
> >
> > E1-(Bound INADDR_ANY)
> >
> > I currently have
> >
> > IP-A: Nic1/IP-B: Nic2
> >
> > I setup an association with a peer
> >
> > A1=(IP-A,IP-B:Port 4444)<====================>(IP-Z:Port 5555)
> >
> > Now everything is fine...
> >
> > Now My administrator plugs in Nic3 and does:
> >
> > ifconfig nic3 IP-C
> >
> > Now ... my endpoint has JUST CHANGED!! It is now
> >
> > IP-A: Nic1/IP-B: Nic2/IP-C: Nic3
> >
> > (remember we are bound to ALL)....
> >
> > Now  my peer does
> >
> >                  <<------TO:IP-C Port4444----INIT(IP:Z port 5555)
> >
> > .....
> >
> > You have the SAME EXACT thing you have been dicussing ..
> >
> >
> > This is one of the reasons we NEED Add-IP.
> >
> > ....If we think about the ADD-IP case... then when the ADDRESS
> > was added to the ENDPOINT (automatically) ... you would have gotten
> >
> > -----ADD-IP(IP-C)-------->
> >
> > and not reached the above condition...
> >
> [ Sandeep ] I agree, that ADD-IP in this case has saved us. But Just a small
> question, What happens that if this endpoint E1 has associations A1 & A2,
> And A1 supports ADD-IP & A2 Does *NOT* support ADD-IP, what should be done
> for A2 after IPc nic was added ?
> 
> I think disconnect A2 and re-establish it again, any other suggestion or
> option ???????
> 
> If this is the case Network administrator should know the consequences(
> disconnecting some associations )of ADDING IP addresses dynamically, Is'nt
> it ?
> 
> > The key is has Qiaobing stated, you ADD addresses to ENDPOINTs. I
> > think we will add some text towards this clearfication in the next
> > pass..
> >
>  [ Sandeep] I am glad that this is being cleared UP, So the same statment is
> TRUE for DELETE-IP also, I.E.  " You DELETE addresses from an Endpoint,
> Although final beneficiary in both the cases is Association, but
> ADDITION/DELETION of IP addresses is done on association."  Correct
> ????????????????????
> 
> If everything said above is TRUE than what i am gonna assume is  " When User
> Does ADDIP i will add that address in all the assocation for that Endpoint,
> & When User does DELETE IP i will delete the address from all the
> association "
> 
> However i have real problem if some association does'nt support ADD/DELETE
> IP procedures then what :-(
> 

Well.. what you are proposing doing is exactly what our kernel work
in BSD's did.. only additional thing is that we do allow a user
to turn this off i.e. .. specify NOT to do the ASCONF.

Note as for the issue when someone does NOT support the ASCONF.. I
don't think you are in ANY worse situation then you were before.
If you really must get the IP address added then the side that
adds the address needs to restart the association...

R

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 15:03:40 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA11417
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 15:03:40 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA13440
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 15:03:24 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA10129;
	Mon, 29 Apr 2002 14:15:15 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA10048
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 14:15:10 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA08437
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 14:15:07 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3TIEoH59994;
	Mon, 29 Apr 2002 13:14:51 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCD8D9A.B9501B76@stewart.chicago.il.us>
Date: Mon, 29 Apr 2002 13:14:50 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
CC: "'bidulock@openss7.org'" <bidulock@openss7.org>,
        Jacob Heitz <jheitz@lucent.com>,
        "Khan, Mohammad" <mo_khan@trillium.com>,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <F92978223B01D311ACFF0008C71EE19D015DF7DC@MCHH225E>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Tuexen Michael wrote:
> 
> Brian,
> 
> see my comments below.
> 
> Best regards
> Michael
> 
> Michael Tuexen    Tel.:   +49 89 722 47210
> Siemens AG        Fax:    +49 89 722 48212
> ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de
> 
> > -----Original Message-----
> > From: Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> > Sent: Monday, April 29, 2002 2:15 PM
> > To:   Tuexen Michael
> > Cc:   Jacob Heitz; Khan, Mohammad; 'tsvwg@ietf.org'; Pillai, Hari
> > Subject:      Re: [Tsvwg] EndPoint and Association Address conflict
> >
> > Tuexen,
> >
> > On Mon, 29 Apr 2002, Tuexen Michael wrote:
> > > >
> > > > Why restrict Add-IP to an endpoint basis?
> > > >
> > >     [Tuexen Michael]  Consider E1 = ({IP'}, Port), E2 = ({IP'''}, Port),
> > >     E3 = ({IP''''}, Port) with pairwise different IP', IP''', IP'''' and two associations
> > >     A1: E1 <-> E2
> > >     A2: E1 <-> E3
> > >
> > >     Now if we add IP'' to the association A1 we get
> > >     A1 : E1' <-> E2
> > >     with E1' = ({IP', IP''}, Port).
> > >     E1' and E2 share one transport address which is in violation to RFC
> > >     2960, page 13.
> >
> > There is only one MUST on page 13: "Two SCTP endpoints MUST NOT have more
> > than one SCTP association between them at any given time."
> >
> > How does this violate anything on page 13?
>         [Tuexen Michael]  I did not mean this sentence.
> > E1' and E2 only have one association between them.  E1 and E3 only have one
> > association between them.  Everthing appears ok as far as page 13 is
> > concerned.
> >
> > Did you mean another page?
> >
> > Or, did you mean the statement: "A transport address used by an
> > SCTP endpoint must not be used by another SCTP endpoint.  In other
> > words, a transport address is unique to an SCTP endpoint."  ...which
> > is absurd, particularly as it is done in TCP all the time, and when
> > you consider that non-multihomed SCTP associations need to reuse
> > dynamic port numbers for connection to other hosts.
>         [Tuexen Michael]  I mean this sentence. This sentence
>         is NOT absurd, at least if you use the definition of
>         SCTP end-point I have given earlier. This is one of the reasons
>         I'm quite sure that the RFC authors had the defintion I gave
>         in mind.
> 
Michael:

Yes this IS the definition we had in mind and as the writer
of add-ip it as always been our intention of having addip be
on an ENDPOINT basis.. as your observations have been...


Regards

R


> > --brian
> >
> > >
> > >     Please note that an association has two SCTP end points and these end
> > >     points have the IP addresses. So changing IP addresses means modifying
> > >     the end points.
> > >
> > >     Best regards
> > >     Michael
> >
> > --
> > Brian F. G. Bidulock    > ¦ The reasonable man adapts himself to the ¦
> > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> >                         ¦ Therefore  all  progress  depends on the ¦
> >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 15:11:57 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA11790
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 15:11:56 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA13851
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 15:11:58 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA12245;
	Mon, 29 Apr 2002 14:42:21 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA12216
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 14:42:18 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA10717
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 14:42:14 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3TIfuH60075;
	Mon, 29 Apr 2002 13:41:56 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCD93F3.CA06B1E@stewart.chicago.il.us>
Date: Mon, 29 Apr 2002 13:41:55 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Qiaobing Xie <qxie1@email.mot.com>
CC: Tuexen Michael <Michael.Tuexen@icn.siemens.de>,
        "'Qiaobing Xie'" <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, s_balani@trillium.com,
        mo_khan@trillium.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F92978223B01D311ACFF0008C71EE19D015DF7D2@MCHH225E> <3CCD857A.7902D0D@email.mot.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Qiaobing Xie wrote:
> 
> Hi, Michael,
> 
> I think two simple rules will be sufficient. Let me use an example to
> illustrate them.
> 
> Assume we have the following endpoints and associations in place:
> 
> on local host         on remote host(s)
> A1: E1 <-----------------> E3
> A2: E1 <---------------------> E4
> A3: E2 <-------------------------> E5
> 
> Rule 1: whenever adding an SCTP transport to a local endpoint, make sure
> this same transport is NOT already present in other existing endpoints
> (e.g., you don't want to add one of E1's transport to E2).
> 
> Rule 2: when adding a transport to a local endpoint that has
> associations to multiple peers (ie, a fan-out config), one MUST send an
> add-ip ASCONF to each of those peers. And the new transport can only be
> used until ALL the peers ack successfully. If anyone of the peers denies
> the add-ip then you can't add it. So, if you add a new transport to E1
> above, you have to send ASCONF to both E3 and E4 and have to get a
> positive ack from both of them.
> 
Q:

I don't agree with this rule.. just because a peer in a UDP model
endpoint on one of its assocaitions is NOT capable of add-ip why
should the rest be stuck. Yes you have an abnormality here in
that that peer COULD send a INIT to the new address.. again I repeat:
1) This is such a corner case it is almost silly.
2) If it happens it is not that hard to handle... Looking at
   the code Peter and I wrote you would end up finding the endpoint
   (since the address WAS added there) and you would then LOOK at
   its associations. This would then appear to be a RESTART where the
   peer added addresses. You would then SEND IT an ABORT saying restart
   while adding addresses. This would leave the other association
un-touched
   and would elimiate the error on the peers part... 


R



> -Qiaobing
> 
> Tuexen Michael wrote:
> >
> > Dear all,
> >
> > looking at the discussion it seems that we have only one restriction
> > for add-IP:
> >
> > Consider the set E of all SCTP end points at a host. Partition
> > E = E_1 union E_2 union ... E_65535
> > into disjoint sets E_P of all end points having port P.
> >
> > When adding an IP-address to multiple end points, all of these
> > end points MUST belong to different E_P.
> >
> > This means that an SCTP implementation has to choose from all end points
> > having the same port on end point to which the add IP procedure is applied.
> >
> > Another point:
> > Consider two associations
> > A1: E1 <-> E2  (The SCTP implementation of E2 is AddIP capable.)
> > A2: E1 <-> E3  (The SCTP implementation of E3 is not AddIP capable.)
> >
> > Now you add an IP address to E1 and get E1'. So A1 is now E1' <-> E2.
> > Then we consider A2 also as A2 : E1' <-> E3, but the stack will never
> > use the added address. Is this correct usage of terminology?
> >
> > Best regards
> > Michael
> >
> > Michael Tuexen    Tel.:   +49 89 722 47210
> > Siemens AG        Fax:    +49 89 722 48212
> > ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de
> >
> > > -----Original Message-----
> > > From: Qiaobing Xie [SMTP:qxie1@email.mot.com]
> > > Sent: Friday, April 26, 2002 2:45 AM
> > > To:   Ivan.Arias-Rodriguez@nokia.com
> > > Cc:   randall@stewart.chicago.il.us; s_balani@trillium.com; Qiaobing Xie; mo_khan@trillium.com; tsvwg@ietf.org
> > > Subject:      Re: [Tsvwg] End point and association Address conflict
> > >
> > > Ivan.Arias-Rodriguez@nokia.com wrote:
> > > ....
> > > >         Now I have another doubt... Imagine that we have an endpoint
> > > > E1 with addresses IPa, IPb and port 1. We have an association A1
> > > > that only uses IPa, whose peer is IPz and port 65535. Thus the
> > > > association A1 can be represented as [(IPa; Port1), (IPz; Port 65535)].
> > > > What happens now if we receive an INIT from IPz/65535 directed to IPb/1?
> > > >
> > > >         As IPb is not being used in A1, we should answer with an
> > > > INIT ACK as in a normal establishment of an association. However,
> > > > when you receive an INIT chunk, you include all the addresses of the endpoint
> > >                                       ^^^^^^^^^^^^^
> > > Here your "..include all the addr..." directly violated the "under no
> > > circumstance can an SCTP transport address ever be present in more than
> > > one SCTP endpoint" rule (see last the paragraph, page 23 of the SCTP
> > > book).
> > >
> > >         E1               E2
> > > A1 = [(IPa; Port1), (IPz; Port 65535)]
> > >
> > >         E3                 E2
> > > A2 = [(IPa, IPb; Port1), (IPz; Port 65535)]
> > >
> > > And you have IPa/Port1 present in both E1 and E3 (they are TWO different
> > > endpoints by definition!). This is prohibited by the SCTP association
> > > definition.
> > >
> > > -Qiaobing
> > >
> > > > in the INIT ACK chunk... But we have to be careful, because in
> > > > this case, when receiving the COOKIE ECHO, we have two options:
> > >
> > >
> > > _______________________________________________
> > > tsvwg mailing list
> > > tsvwg@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 15:25:50 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA13551
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 15:25:50 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA14658
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 15:25:52 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA12678;
	Mon, 29 Apr 2002 14:53:52 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA12650
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 14:53:49 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA11152
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 14:53:43 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3TIrWH60118;
	Mon, 29 Apr 2002 13:53:33 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCD96AC.A1397384@stewart.chicago.il.us>
Date: Mon, 29 Apr 2002 13:53:32 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Qiaobing Xie <qxie1@email.mot.com>
CC: "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <A1171C2850D8D4119D7B0002A508C764016983C9@bgsmsx91.iind.intel.com> <3CCD8BC8.CB90FA61@stewart.chicago.il.us> <3CCD94DF.E9AC7D89@email.mot.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Qiaobing:

See my response to your email on "new rules" I think
I show that all Hell will NOT break loose... since
the handling of duplicate INIT's with the added
"no new addresses can be added (in the implementors guide)"
pretty much handles this case correctly..

R

Qiaobing Xie wrote:
> 
> Randall Stewart wrote:
> >
> > "Khan, Mohammad" wrote:
> > > As per the assumption 2) to add an IPZ at E1 SCTP has to send ASCONF chunk
> > > on all the A1, A2 and A3 associations and it shall not consider this address
> > > as added till it receives success from all the associations. If failure is
> > > received is received from any of the associations then this Add-IP request
> > > shall be discarded.
> > >
> >
> > Here I disagree.. You add it to the endpoint and it is added
> > RIGHT away..
> >
> > At the same time you start sending off your ASCONF's. If for some reason
> > an association  fails to add an address (either via non-support of
> > add-ip OR some other failure) that association is in a odd state
> > until either:
> >
> > A) it restarts
> >
> > or
> >
> > B) it terminates.
> 
> Either treatment (ie., to reject the add or to force out associations
> that can't perform the add) will be OK with me. The bottom line is to
> maintain the integrity of the endpoint, otherwise "all hell will break
> loose" :-)
> 
> -Qiaobing
> 
> >
> > This is such a corner case I don't see it as an issue. The corner
> > case being that your peer (who does NOT support Add-IP) sends an
> > INIT to the new address. If that were to occur, when you unpacked
> > the cookie you would find the other association and in my opinion
> > to handle the corner case.. abort the newly forming association
> > (not the old one) or just drop the cookie and let it timeout until
> > it goes away...
> >
> > I don't see this to be an issue...
> >
> > R
> > --
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 15:31:08 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA14514
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 15:31:03 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA15122
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 15:31:05 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA12296;
	Mon, 29 Apr 2002 14:43:07 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA11921
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 14:36:08 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA10548
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 14:36:04 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3TIZJH60044;
	Mon, 29 Apr 2002 13:35:19 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCD9267.6456A90E@stewart.chicago.il.us>
Date: Mon, 29 Apr 2002 13:35:19 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Jon Grimm <jgrimm@austin.ibm.com>
CC: Ivan.Arias-Rodriguez@nokia.com, bidulock@openss7.org,
        Michael.Tuexen@icn.siemens.de, qxie1@email.mot.com,
        s_balani@trillium.com, Qiaobing.Xie@motorola.com, mo_khan@trillium.com,
        h_pillai@trillium.com, jheitz@lucent.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFB4@esebe017.NOE.Nokia.com> <3CCD844A.DB1BC960@austin.ibm.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Jon:

a couple of comments :>

Jon Grimm wrote:
> 
> Ivan/All
> 
>         I think something that needs to come out of this discussion are
> scenarios and mandatory behavior in those scenarios (or why flexibility
> is allowed if not mandatory behavior).  Its a little disturbing that a
> similar 'what is an endpoint?' discussion seems to pop up again every so
> often.  (e.g.bindx discussion last Sept in
> ftp://ftp.ietf.org/ietf-mail-archive/tsvwg/2001-09.mail ) and still does
> not seem to be completely clear even to those following SCTP closely
> over the last year.


I agree.. it is a discussion that seems to NEVER end :>


> 
> For each of the following scenarios,
> 
> 1) A new address is to be added an endpoint. (manual)
> 2) An address is to be deleted from an endpoint.  (manual)
> 3) A new NIC/address is configured and an endpoint is bound to all
> (auto)
> 4) An address bound is brought down (no longer available) (auto).
> 5) An address bound is brought down for maintenance and then brought
> on-line again. (auto)
> 
> consider each of the following subcases (assuming multiple associations
> under an ep).
> 
> a) All Peers have ADD-IP.
> b) No Peers have ADD-IP.
> c) A subset of Peers support ADD-IP.
> 
> What happens?
> 


Where are you proposing putting these examples? I think some
text is needed for clearification but I am not sure that a protocol
spec is where you illustrate all these different examples...

> Another comment:
> Dropping down a little bit into the Sockets API for illustration:
> 
> Depending on one's interpretation of endpoint, what is the endpoint for
> the association of a peeled-off UDP style socket or for that matter an
> 'accepted' TCP-style socket.  If the strict interpretation of the
> endpoint is kept.. wouldn't an ADD-IP on the child sockets (peeled off
> or accept TCP) affect the local endpoint of the parent socket?  Or vise
> versa, should an ADD-IP on a TCP-style listening socket, affect the
> accepted sockets?  If yes, was this expected?  If no, does the original

In my view the ADD-IP is done against the listening type endpoint. So
in the case of connected and/or peeled off associations they are
a different critter. In some ways they have formed a "new" endpoint. But
a interesting thing about them is they CAN NOT listen. So maybe in
this case Brian's adding an address to the association is the right
approach. They can NEVER receive a new association.. only the single
listner (in the TCP or UDP case) can do that. 

So a peeled off or "connected" socket with only one association may
well be a candidate to do an add-ip on an association basis.. but not
on the listening endpoint or a UDP model grouping of associations...


> definition of endpoint still hold?
> 
> Even if each expected behavior does not get spelled it, it would be nice
> for the SCTP Sockets API I-D to clarify the relationship between a
> socket and an endpoint, so implementors can at least infer what the
> intended behavior should be.

Hmm this may be the best place for this information since it will
be informational..

R

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 15:32:12 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA14626
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 15:32:12 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA15328
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 15:32:14 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA12449;
	Mon, 29 Apr 2002 14:45:16 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA12416
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 14:45:13 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA10896
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 14:45:08 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3TIiTH60087;
	Mon, 29 Apr 2002 13:44:29 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCD948C.A1E11EBC@stewart.chicago.il.us>
Date: Mon, 29 Apr 2002 13:44:29 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Qiaobing Xie <qxie1@email.mot.com>
CC: Ivan.Arias-Rodriguez@nokia.com, bidulock@openss7.org,
        Michael.Tuexen@icn.siemens.de,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>, s_balani@trillium.com,
        mo_khan@trillium.com, h_pillai@trillium.com, jheitz@lucent.com,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFB4@esebe017.NOE.Nokia.com> <3CCD8A95.72572C78@email.mot.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Qiaobing Xie wrote:
> 
> Hi, Ivan:
> 
> Ivan.Arias-Rodriguez@nokia.com wrote:
> ....snip....
> >
> >         Summarizing:
> >
> >         - An endpoint can use a subset of the source addresses of the host.
> 
> Yes.
> 
> >         - An association, can use a subset of the source addresses of the
> >           endpoint to which it belongs.
> 
> No, this is incorrect.
> 

Sorry it IS correct. Forget add-ip.. if you say the above then
a basic bound-all endpoint begins to violate this new "rule"
when you plug in a new card and ifconfig it up.

There is no reason that an endpoint can't have associations that
end up using a sub-set of the endpoints addresses. And I have
shown in my previous email what will happen if this endpoints
silly peer attempts to INIT to the NEW address... It all
works out..

R

> All associations coming out of an endpoint MUST use all the source
> addresses of the endpoint all the time. However, the address set of the
> endpoint can change over time (by adding or deleting IPs). Whenever that
> happens ALL the existing associations coming out of the endpoint MUST be
> updated to the new address set (this is why we have ASCONF exchanges).
> Moreover, there are rules over how the address set of the endpoint can
> be changed. That is what I sent out in my previous email.
> 
> regards,
> -Qiaobing
> 
> >
> >         I would very much appreciate comments about this mail...
> >
> >         Thanks a lot!
> >
> >         BR Iván Arias Rodríguez

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 16:27:41 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA18531
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 16:27:40 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA19285
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 16:27:43 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA16457;
	Mon, 29 Apr 2002 15:59:39 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA16419
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 15:59:36 -0400 (EDT)
Received: from ilsa.franken.de (postfix@ilsa.franken.de [193.175.24.42])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA17389
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 15:59:31 -0400 (EDT)
Received: from localhost (pD9E4F8FD.dip.t-dialin.net [217.228.248.253])
	by ilsa.franken.de (Postfix) with ESMTP
	id 74E62245FD; Mon, 29 Apr 2002 21:59:10 +0200 (CEST)
Date: Mon, 29 Apr 2002 21:54:59 +0200
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Mime-Version: 1.0 (Apple Message framework v481)
Cc: Qiaobing Xie <qxie1@email.mot.com>, Ivan.Arias-Rodriguez@nokia.com,
        bidulock@openss7.org, Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        s_balani@trillium.com, mo_khan@trillium.com, h_pillai@trillium.com,
        jheitz@lucent.com, tsvwg@ietf.org
To: Randall Stewart <randall@stewart.chicago.il.us>
From: Michael Tuexen <Michael.Tuexen@icn.siemens.de>
In-Reply-To: <3CCD948C.A1E11EBC@stewart.chicago.il.us>
Message-Id: <FC9160EC-5BAA-11D6-ACAA-0030654C1AB6@icn.siemens.de>
X-Mailer: Apple Mail (2.481)
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id PAA16422
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Dear all,

since we agree on the end-point terminology we can go further.

See my comment below.

Best regards
Michael

On Monday, April 29, 2002, at 08:44 PM, Randall Stewart wrote:

> Qiaobing Xie wrote:
>>
>> Hi, Ivan:
>>
>> Ivan.Arias-Rodriguez@nokia.com wrote:
>> ....snip....
>>>
>>>         Summarizing:
>>>
>>>         - An endpoint can use a subset of the source addresses of the 
>>> host.
>>
>> Yes.
>>
>>>         - An association, can use a subset of the source addresses of 
>>> the
>>>           endpoint to which it belongs.
>>
>> No, this is incorrect.
>>
>
> Sorry it IS correct. Forget add-ip.. if you say the above then
> a basic bound-all endpoint begins to violate this new "rule"
> when you plug in a new card and ifconfig it up.
>
> There is no reason that an endpoint can't have associations that
> end up using a sub-set of the endpoints addresses. And I have
> shown in my previous email what will happen if this endpoints
> silly peer attempts to INIT to the NEW address... It all
> works out..
>
> R
I'm not sure what the term 'an association can use a subset of IP 
addresses' means.
I'm interpreting it as 'use the address in an INIT or INIT-ACK chunk'. 
Taking that
definition, I think, the address scoping of IPv6 (and IPv4) will reduce 
the set
of used IP addresses possibly to a proper subset of the address of the 
end-point.
Is this correct?
>
>> All associations coming out of an endpoint MUST use all the source
>> addresses of the endpoint all the time. However, the address set of the
>> endpoint can change over time (by adding or deleting IPs). Whenever 
>> that
>> happens ALL the existing associations coming out of the endpoint MUST 
>> be
>> updated to the new address set (this is why we have ASCONF exchanges).
>> Moreover, there are rules over how the address set of the endpoint can
>> be changed. That is what I sent out in my previous email.
>>
>> regards,
>> -Qiaobing
>>
>>>
>>>         I would very much appreciate comments about this mail...
>>>
>>>         Thanks a lot!
>>>
>>>         BR Iván Arias Rodríguez
>
> --
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
>
Michael.Tuexen@micmac.franken.de


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 16:28:11 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA18588
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 16:28:11 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA19313
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 16:28:13 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA16266;
	Mon, 29 Apr 2002 15:55:21 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA16237
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 15:55:18 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-230.dallas.net [209.44.42.230])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA17201
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 15:55:12 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id OAA28950;
	Mon, 29 Apr 2002 14:54:46 -0500
Date: Mon, 29 Apr 2002 14:54:46 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Randall Stewart <randall@stewart.chicago.il.us>
Cc: Qiaobing Xie <qxie1@email.mot.com>,
        Tuexen Michael <Michael.Tuexen@icn.siemens.de>,
        "'Qiaobing Xie'" <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, s_balani@trillium.com,
        mo_khan@trillium.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
Message-ID: <20020429145446.G15434@openss7.org>
Reply-To: bidulock@openss7.org
References: <F92978223B01D311ACFF0008C71EE19D015DF7D2@MCHH225E> <3CCD857A.7902D0D@email.mot.com> <3CCD93F3.CA06B1E@stewart.chicago.il.us>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <3CCD93F3.CA06B1E@stewart.chicago.il.us>; from randall@stewart.chicago.il.us on Mon, Apr 29, 2002 at 01:41:55PM -0500
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Randall,

On Mon, 29 Apr 2002, Randall Stewart wrote:

> Qiaobing Xie wrote:
> > 
> > Hi, Michael,
> > 
> > I think two simple rules will be sufficient. Let me use an example to
> > illustrate them.
> > 
> > Assume we have the following endpoints and associations in place:
> > 
> > on local host         on remote host(s)
> > A1: E1 <-----------------> E3
> > A2: E1 <---------------------> E4
> > A3: E2 <-------------------------> E5
> > 
> > Rule 1: whenever adding an SCTP transport to a local endpoint, make sure
> > this same transport is NOT already present in other existing endpoints
> > (e.g., you don't want to add one of E1's transport to E2).
> > 
> > Rule 2: when adding a transport to a local endpoint that has
> > associations to multiple peers (ie, a fan-out config), one MUST send an
> > add-ip ASCONF to each of those peers. And the new transport can only be
> > used until ALL the peers ack successfully. If anyone of the peers denies
> > the add-ip then you can't add it. So, if you add a new transport to E1
> > above, you have to send ASCONF to both E3 and E4 and have to get a
> > positive ack from both of them.
> > 
> Q:
> 
> I don't agree with this rule.. just because a peer in a UDP model
> endpoint on one of its assocaitions is NOT capable of add-ip why
> should the rest be stuck. Yes you have an abnormality here in
> that that peer COULD send a INIT to the new address.. again I repeat:
> 1) This is such a corner case it is almost silly.
> 2) If it happens it is not that hard to handle... Looking at
>    the code Peter and I wrote you would end up finding the endpoint
>    (since the address WAS added there) and you would then LOOK at
>    its associations. This would then appear to be a RESTART where the
>    peer added addresses. You would then SEND IT an ABORT saying restart
>    while adding addresses. This would leave the other association
> un-touched
>    and would elimiate the error on the peers part... 
> 

I disagree with Rule 2 as well, but for different reasons: there may be
other reasons for not wanting to add the IP to an existing association
with, say, E1.

--brian


-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 16:41:44 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA19123
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 16:41:44 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA20335
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 16:41:47 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA18457;
	Mon, 29 Apr 2002 16:17:05 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA17531
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 16:07:34 -0400 (EDT)
Received: from mg03.austin.ibm.com (mg03.austin.ibm.com [192.35.232.20])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA17769
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 16:07:31 -0400 (EDT)
Received: from austin.ibm.com (netmail1.austin.ibm.com [9.3.7.138])
	by mg03.austin.ibm.com (AIX4.3/8.9.3/8.9.3) with ESMTP id PAA20312;
	Mon, 29 Apr 2002 15:03:07 -0500
Received: from austin.ibm.com (sctptest.dynamic.austin.ibm.com [9.111.216.144] (may be forged))
	by austin.ibm.com (AIX4.3/8.9.3/8.9.3) with ESMTP id PAA30066;
	Mon, 29 Apr 2002 15:07:28 -0500
Message-ID: <3CCDA2AD.CB11491@austin.ibm.com>
Date: Mon, 29 Apr 2002 14:44:45 -0500
From: Jon Grimm <jgrimm@austin.ibm.com>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.4.18lksctp i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Randall Stewart <randall@stewart.chicago.il.us>
CC: Ivan.Arias-Rodriguez@nokia.com, bidulock@openss7.org,
        Michael.Tuexen@icn.siemens.de, qxie1@email.mot.com,
        s_balani@trillium.com, Qiaobing.Xie@motorola.com, mo_khan@trillium.com,
        h_pillai@trillium.com, jheitz@lucent.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFB4@esebe017.NOE.Nokia.com> <3CCD844A.DB1BC960@austin.ibm.com> <3CCD9267.6456A90E@stewart.chicago.il.us>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Thanks for the remarks, Randy. 
 I have a few comments still below.  Thanks, Jon

Randall Stewart wrote:
> 
> Jon:
> 
> a couple of comments :>
> 
> Jon Grimm wrote:
> >
> > Ivan/All
> >
> >         I think something that needs to come out of this discussion are
> > scenarios and mandatory behavior in those scenarios (or why flexibility
> > is allowed if not mandatory behavior).  Its a little disturbing that a
> > similar 'what is an endpoint?' discussion seems to pop up again every so
> > often.  (e.g.bindx discussion last Sept in
> > ftp://ftp.ietf.org/ietf-mail-archive/tsvwg/2001-09.mail ) and still does
> > not seem to be completely clear even to those following SCTP closely
> > over the last year.
> 
> I agree.. it is a discussion that seems to NEVER end :>
> 
> >
> > For each of the following scenarios,
> >
> > 1) A new address is to be added an endpoint. (manual)
> > 2) An address is to be deleted from an endpoint.  (manual)
> > 3) A new NIC/address is configured and an endpoint is bound to all
> > (auto)
> > 4) An address bound is brought down (no longer available) (auto).
> > 5) An address bound is brought down for maintenance and then brought
> > on-line again. (auto)
> >
> > consider each of the following subcases (assuming multiple associations
> > under an ep).
> >
> > a) All Peers have ADD-IP.
> > b) No Peers have ADD-IP.
> > c) A subset of Peers support ADD-IP.
> >
> > What happens?
> >
> 
> Where are you proposing putting these examples? I think some
> text is needed for clearification but I am not sure that a protocol
> spec is where you illustrate all these different examples...


In some I-D,of course.   I see you are considering the Sockets API I-D,
yes this could work for me as an appendix there.  Given this latest
round of discussion, folks may wildly disagree, but I'm coming from the
angle that I just want it captured.  ;-)

> 
> > Another comment:
> > Dropping down a little bit into the Sockets API for illustration:
> >
> > Depending on one's interpretation of endpoint, what is the endpoint for
> > the association of a peeled-off UDP style socket or for that matter an
> > 'accepted' TCP-style socket.  If the strict interpretation of the
> > endpoint is kept.. wouldn't an ADD-IP on the child sockets (peeled off
> > or accept TCP) affect the local endpoint of the parent socket?  Or vise
> > versa, should an ADD-IP on a TCP-style listening socket, affect the
> > accepted sockets?  If yes, was this expected?  If no, does the original
> 
> In my view the ADD-IP is done against the listening type endpoint. So
> in the case of connected and/or peeled off associations they are
> a different critter. In some ways they have formed a "new" endpoint. But
> a interesting thing about them is they CAN NOT listen. So maybe in
> this case Brian's adding an address to the association is the right
> approach. They can NEVER receive a new association.. only the single
> listner (in the TCP or UDP case) can do that.
> 
> So a peeled off or "connected" socket with only one association may
> well be a candidate to do an add-ip on an association basis.. but not
> on the listening endpoint or a UDP model grouping of associations...
> 


Has this gone full circle then?   This sounds more like you are always
doing add-ip on associations.   Or you are really always doing add-ip on
(Brian's or Ivan's ??) individual communication endpoints that belong to
a grouping (group of 1 (peeled off) or many (udp style))?  

Consider two cases:    
1)  If not bound-all, and an address is added via bindx.. I'd want the
bindx to apply to the listening endpoint, as well as the associations
currently grouped under that socket, but not the peeled-off socket.   
2) If however, if the original socket was bound-all and a new interface
is configured, I'd want to this new information to propogate both to the
original listening socket and any peeled off sockets.   

Do these make sense?


A related question:  should Add-ip happen on associations waiting in the
TCP-style accept queue?   manual bindx()?  auto INADDR_ANY??



> > definition of endpoint still hold?
> >
> > Even if each expected behavior does not get spelled it, it would be nice
> > for the SCTP Sockets API I-D to clarify the relationship between a
> > socket and an endpoint, so implementors can at least infer what the
> > intended behavior should be.
> 
> Hmm this may be the best place for this information since it will
> be informational..

Sure.  If thats were folks want it, as long as gets agreed on and
captured somewhere.  

Thanks,
Jon


> 
> R
> 
> --
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 16:46:18 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA19220
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 16:46:17 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA20469
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 16:46:20 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA18275;
	Mon, 29 Apr 2002 16:16:38 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA12326
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 14:43:10 -0400 (EDT)
Received: from motgate2.mot.com (motgate2.mot.com [136.182.1.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA10729
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 14:43:08 -0400 (EDT)
Received: [from pobox.mot.com (pobox.mot.com [129.188.137.100]) by motgate2.mot.com (motgate2 2.1) with ESMTP id LAA24581; Mon, 29 Apr 2002 11:43:00 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox.mot.com (MOT-pobox 2.0) with ESMTP id LAA16008; Mon, 29 Apr 2002 11:43:00 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3TIgsb04926;
	Mon, 29 Apr 2002 13:42:54 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id NAA21766;
	Mon, 29 Apr 2002 13:43:04 -0500 (CDT)
Message-ID: <3CCD94DF.E9AC7D89@email.mot.com>
Date: Mon, 29 Apr 2002 13:45:51 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Randall Stewart <randall@stewart.chicago.il.us>
CC: "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org,
        "Pillai, Hari" <h_pillai@trillium.com>
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <A1171C2850D8D4119D7B0002A508C764016983C9@bgsmsx91.iind.intel.com> <3CCD8BC8.CB90FA61@stewart.chicago.il.us>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Randall Stewart wrote:
> 
> "Khan, Mohammad" wrote:
> > As per the assumption 2) to add an IPZ at E1 SCTP has to send ASCONF chunk
> > on all the A1, A2 and A3 associations and it shall not consider this address
> > as added till it receives success from all the associations. If failure is
> > received is received from any of the associations then this Add-IP request
> > shall be discarded.
> >
> 
> Here I disagree.. You add it to the endpoint and it is added
> RIGHT away..
> 
> At the same time you start sending off your ASCONF's. If for some reason
> an association  fails to add an address (either via non-support of
> add-ip OR some other failure) that association is in a odd state
> until either:
> 
> A) it restarts
> 
> or
> 
> B) it terminates.

Either treatment (ie., to reject the add or to force out associations
that can't perform the add) will be OK with me. The bottom line is to
maintain the integrity of the endpoint, otherwise "all hell will break
loose" :-)

-Qiaobing


> 
> This is such a corner case I don't see it as an issue. The corner
> case being that your peer (who does NOT support Add-IP) sends an
> INIT to the new address. If that were to occur, when you unpacked
> the cookie you would find the other association and in my opinion
> to handle the corner case.. abort the newly forming association
> (not the old one) or just drop the cookie and let it timeout until
> it goes away...
> 
> I don't see this to be an issue...
> 
> R
> --
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 16:49:59 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA19339
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 16:49:58 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA20652
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 16:50:01 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA18384;
	Mon, 29 Apr 2002 16:16:53 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA14167
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 15:16:24 -0400 (EDT)
Received: from motgate.mot.com (motgate.mot.com [129.188.136.100])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA12060
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 15:16:21 -0400 (EDT)
Received: [from pobox4.mot.com (pobox4.mot.com [10.64.251.243]) by motgate.mot.com (motgate 2.1) with ESMTP id MAA23416; Mon, 29 Apr 2002 12:16:23 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox4.mot.com (MOT-pobox4 2.0) with ESMTP id MAA09201; Mon, 29 Apr 2002 12:16:22 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3TJGIb07998;
	Mon, 29 Apr 2002 14:16:18 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id OAA21788;
	Mon, 29 Apr 2002 14:16:28 -0500 (CDT)
Message-ID: <3CCD9CB4.E3A1B91C@email.mot.com>
Date: Mon, 29 Apr 2002 14:19:16 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Randall Stewart <randall@stewart.chicago.il.us>
CC: Qiaobing Xie <Qiaobing.Xie@motorola.com>, Ivan.Arias-Rodriguez@nokia.com,
        bidulock@openss7.org, Michael.Tuexen@icn.siemens.de,
        s_balani@trillium.com, mo_khan@trillium.com, h_pillai@trillium.com,
        jheitz@lucent.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFB4@esebe017.NOE.Nokia.com> <3CCD8A95.72572C78@email.mot.com> <3CCD948C.A1E11EBC@stewart.chicago.il.us>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

...
> > No, this is incorrect.
> >
> 
> Sorry it IS correct. Forget add-ip.. if you say the above then
> a basic bound-all endpoint begins to violate this new "rule"
> when you plug in a new card and ifconfig it up.

Randy,

See if I understand what you are trying to say here. Say, you have the
following "bound-all" endpoint with an assoc to a peer E2

host has: IPA and IPB
     E1             A1     E2
[IPA,IPB:Port1] <-------> [...]  /* you use all addr in E1 */

you then plug a new nic and ifconfig it up with IPC, 

host has: IPA, IPB, and IPC
     E1             A1     E2
[IPA,IPB:Port1] <-------> [...]  /* you use still all addr in E1 */

you attempt to update E1 but add-ip procedure and has succeeded

host has: IPA, IPB, and IPC
     E1'                A1'    E2
[IPA,IPB,IPC:Port1] <-------> [...]  /* you use all addr in E1' */

If add-ip did go through

host has: IPA, IPB, and IPC
     E1             A1     E2
[IPA,IPB:Port1] <-------> [...]  /* you use still all addr in E1 */

So, I don't see any conflict between the rule and your case. 

> 
> There is no reason that an endpoint can't have associations that
> end up using a sub-set of the endpoints addresses. And I have
> shown in my previous email what will happen if this endpoints
> silly peer attempts to INIT to the NEW address... It all
> works out..

I agree with you that forcing the silly peer out and to restart is a
good alternative in this corner case.

-Qiaobing

> 
> R
> 
> > All associations coming out of an endpoint MUST use all the source
> > addresses of the endpoint all the time. However, the address set of the
> > endpoint can change over time (by adding or deleting IPs). Whenever that
> > happens ALL the existing associations coming out of the endpoint MUST be
> > updated to the new address set (this is why we have ASCONF exchanges).
> > Moreover, there are rules over how the address set of the endpoint can
> > be changed. That is what I sent out in my previous email.
> >
> > regards,
> > -Qiaobing
> >
> > >
> > >         I would very much appreciate comments about this mail...
> > >
> > >         Thanks a lot!
> > >
> > >         BR Iván Arias Rodríguez
> 
> --
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 16:51:28 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA19446
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 16:51:28 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA20749
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 16:51:31 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA18997;
	Mon, 29 Apr 2002 16:23:19 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA18969
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 16:23:17 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA18350
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 16:23:12 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3TKMiH60306;
	Mon, 29 Apr 2002 15:22:44 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCDAB94.5A70AFAF@stewart.chicago.il.us>
Date: Mon, 29 Apr 2002 15:22:44 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Qiaobing Xie <qxie1@email.mot.com>
CC: Qiaobing Xie <Qiaobing.Xie@motorola.com>, Ivan.Arias-Rodriguez@nokia.com,
        bidulock@openss7.org, Michael.Tuexen@icn.siemens.de,
        s_balani@trillium.com, mo_khan@trillium.com, h_pillai@trillium.com,
        jheitz@lucent.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFB4@esebe017.NOE.Nokia.com> <3CCD8A95.72572C78@email.mot.com> <3CCD948C.A1E11EBC@stewart.chicago.il.us> <3CCD9CB4.E3A1B91C@email.mot.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Qiaobing Xie wrote:
> 
> ...
> > > No, this is incorrect.
> > >
> >
> > Sorry it IS correct. Forget add-ip.. if you say the above then
> > a basic bound-all endpoint begins to violate this new "rule"
> > when you plug in a new card and ifconfig it up.
> 
> Randy,
> 
> See if I understand what you are trying to say here. Say, you have the
> following "bound-all" endpoint with an assoc to a peer E2
> 
> host has: IPA and IPB
>      E1             A1     E2
> [IPA,IPB:Port1] <-------> [...]  /* you use all addr in E1 */
> 
> you then plug a new nic and ifconfig it up with IPC,
> 
> host has: IPA, IPB, and IPC
>      E1             A1     E2
> [IPA,IPB:Port1] <-------> [...]  /* you use still all addr in E1 */
> 
> you attempt to update E1 but add-ip procedure and has succeeded
> 
> host has: IPA, IPB, and IPC
>      E1'                A1'    E2
> [IPA,IPB,IPC:Port1] <-------> [...]  /* you use all addr in E1' */
> 
> If add-ip did go through
> 
> host has: IPA, IPB, and IPC
>      E1             A1     E2
> [IPA,IPB:Port1] <-------> [...]  /* you use still all addr in E1 */
> 
> So, I don't see any conflict between the rule and your case.
> 
> >
> > There is no reason that an endpoint can't have associations that
> > end up using a sub-set of the endpoints addresses. And I have
> > shown in my previous email what will happen if this endpoints
> > silly peer attempts to INIT to the NEW address... It all
> > works out..
> 
> I agree with you that forcing the silly peer out and to restart is a
> good alternative in this corner case.
> 
Thinking about this with a private email discussion I realize
the address add restriction is the peers address... maybe we
need to expand that to the local addresses too. this way 
the restart routine would catch it.. abort the silly peer and
we don't have to worry about this case..

R


> -Qiaobing
> 
> >
> > R
> >
> > > All associations coming out of an endpoint MUST use all the source
> > > addresses of the endpoint all the time. However, the address set of the
> > > endpoint can change over time (by adding or deleting IPs). Whenever that
> > > happens ALL the existing associations coming out of the endpoint MUST be
> > > updated to the new address set (this is why we have ASCONF exchanges).
> > > Moreover, there are rules over how the address set of the endpoint can
> > > be changed. That is what I sent out in my previous email.
> > >
> > > regards,
> > > -Qiaobing
> > >
> > > >
> > > >         I would very much appreciate comments about this mail...
> > > >
> > > >         Thanks a lot!
> > > >
> > > >         BR Iván Arias Rodríguez
> >
> > --
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 16:58:13 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA19610
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 16:58:09 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA20951
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 16:58:12 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA17752;
	Mon, 29 Apr 2002 16:10:18 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA17724
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 16:10:16 -0400 (EDT)
Received: from ilsa.franken.de (postfix@ilsa.franken.de [193.175.24.42])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA17854
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 16:10:12 -0400 (EDT)
Received: from localhost (pD9E4F8FD.dip.t-dialin.net [217.228.248.253])
	by ilsa.franken.de (Postfix) with ESMTP
	id F1FF2245FD; Mon, 29 Apr 2002 22:10:12 +0200 (CEST)
Date: Mon, 29 Apr 2002 22:06:01 +0200
Subject: Re: [Tsvwg] End point and association Address conflict
Content-Type: text/plain; charset=US-ASCII; format=flowed
Mime-Version: 1.0 (Apple Message framework v481)
Cc: "'Qiaobing Xie'" <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, randall@stewart.chicago.il.us,
        s_balani@trillium.com, mo_khan@trillium.com, tsvwg@ietf.org
To: Qiaobing Xie <qxie1@email.mot.com>
From: Michael Tuexen <Michael.Tuexen@icn.siemens.de>
In-Reply-To: <3CCD857A.7902D0D@email.mot.com>
Message-Id: <875228B9-5BAC-11D6-ACAA-0030654C1AB6@icn.siemens.de>
Content-Transfer-Encoding: 7bit
X-Mailer: Apple Mail (2.481)
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi Qiaobing,

see my comments below.

Best regards
Michael

On Monday, April 29, 2002, at 07:40 PM, Qiaobing Xie wrote:

> Hi, Michael,
>
> I think two simple rules will be sufficient. Let me use an example to
> illustrate them.
>
> Assume we have the following endpoints and associations in place:
>
> on local host         on remote host(s)
> A1: E1 <-----------------> E3
> A2: E1 <---------------------> E4
> A3: E2 <-------------------------> E5
>
> Rule 1: whenever adding an SCTP transport to a local endpoint, make sure
> this same transport is NOT already present in other existing endpoints
> (e.g., you don't want to add one of E1's transport to E2).
>
> Rule 2: when adding a transport to a local endpoint that has
> associations to multiple peers (ie, a fan-out config), one MUST send an
> add-ip ASCONF to each of those peers. And the new transport can only be
> used until ALL the peers ack successfully. If anyone of the peers denies
> the add-ip then you can't add it. So, if you add a new transport to E1
> above, you have to send ASCONF to both E3 and E4 and have to get a
> positive ack from both of them.
>
I do NOT like this rule. How do you handle the case if E3 and E4 support 
AddIP, but
E5 does not. Do you say that you can not add an address to E1 because 
one of the peers
does not support AddIP? for deployment this means that AddIP only be 
used if all
implementations support this...
Why not add it to E1, but it is not allowed to be used as a source 
address for A5.
> -Qiaobing
>
> Tuexen Michael wrote:
>>
>> Dear all,
>>
>> looking at the discussion it seems that we have only one restriction
>> for add-IP:
>>
>> Consider the set E of all SCTP end points at a host. Partition
>> E = E_1 union E_2 union ... E_65535
>> into disjoint sets E_P of all end points having port P.
>>
>> When adding an IP-address to multiple end points, all of these
>> end points MUST belong to different E_P.
>>
>> This means that an SCTP implementation has to choose from all end 
>> points
>> having the same port on end point to which the add IP procedure is 
>> applied.
>>
>> Another point:
>> Consider two associations
>> A1: E1 <-> E2  (The SCTP implementation of E2 is AddIP capable.)
>> A2: E1 <-> E3  (The SCTP implementation of E3 is not AddIP capable.)
>>
>> Now you add an IP address to E1 and get E1'. So A1 is now E1' <-> E2.
>> Then we consider A2 also as A2 : E1' <-> E3, but the stack will never
>> use the added address. Is this correct usage of terminology?
>>
>> Best regards
>> Michael
>>
>> Michael Tuexen    Tel.:   +49 89 722 47210
>> Siemens AG        Fax:    +49 89 722 48212
>> ICN WN CC SE 7    E-mail: Michael.Tuexen@icn.siemens.de
>>
>>> -----Original Message-----
>>> From: Qiaobing Xie [SMTP:qxie1@email.mot.com]
>>> Sent: Friday, April 26, 2002 2:45 AM
>>> To:   Ivan.Arias-Rodriguez@nokia.com
>>> Cc:   randall@stewart.chicago.il.us; s_balani@trillium.com; Qiaobing 
>>> Xie; mo_khan@trillium.com; tsvwg@ietf.org
>>> Subject:      Re: [Tsvwg] End point and association Address conflict
>>>
>>> Ivan.Arias-Rodriguez@nokia.com wrote:
>>> ....
>>>>         Now I have another doubt... Imagine that we have an endpoint
>>>> E1 with addresses IPa, IPb and port 1. We have an association A1
>>>> that only uses IPa, whose peer is IPz and port 65535. Thus the
>>>> association A1 can be represented as [(IPa; Port1), (IPz; Port 
>>>> 65535)].
>>>> What happens now if we receive an INIT from IPz/65535 directed to 
>>>> IPb/1?
>>>>
>>>>         As IPb is not being used in A1, we should answer with an
>>>> INIT ACK as in a normal establishment of an association. However,
>>>> when you receive an INIT chunk, you include all the addresses of the 
>>>> endpoint
>>>                                       ^^^^^^^^^^^^^
>>> Here your "..include all the addr..." directly violated the "under no
>>> circumstance can an SCTP transport address ever be present in more 
>>> than
>>> one SCTP endpoint" rule (see last the paragraph, page 23 of the SCTP
>>> book).
>>>
>>>         E1               E2
>>> A1 = [(IPa; Port1), (IPz; Port 65535)]
>>>
>>>         E3                 E2
>>> A2 = [(IPa, IPb; Port1), (IPz; Port 65535)]
>>>
>>> And you have IPa/Port1 present in both E1 and E3 (they are TWO 
>>> different
>>> endpoints by definition!). This is prohibited by the SCTP association
>>> definition.
>>>
>>> -Qiaobing
>>>
>>>> in the INIT ACK chunk... But we have to be careful, because in
>>>> this case, when receiving the COOKIE ECHO, we have two options:
>>>
>>>
>>> _______________________________________________
>>> tsvwg mailing list
>>> tsvwg@ietf.org
>>> https://www1.ietf.org/mailman/listinfo/tsvwg
>>>
Michael.Tuexen@micmac.franken.de


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 17:01:57 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA19781
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 17:01:57 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA22049
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 17:02:00 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA18662;
	Mon, 29 Apr 2002 16:18:25 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA18125
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 16:15:18 -0400 (EDT)
Received: from ilsa.franken.de (postfix@ilsa.franken.de [193.175.24.42])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA18087
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 16:15:15 -0400 (EDT)
Received: from localhost (pD9E4F8FD.dip.t-dialin.net [217.228.248.253])
	by ilsa.franken.de (Postfix) with ESMTP
	id 34643245FD; Mon, 29 Apr 2002 22:15:14 +0200 (CEST)
Date: Mon, 29 Apr 2002 22:11:01 +0200
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
Content-Type: text/plain; charset=US-ASCII; format=flowed
Mime-Version: 1.0 (Apple Message framework v481)
Cc: Randall Stewart <randall@stewart.chicago.il.us>,
        Ivan.Arias-Rodriguez@nokia.com, bidulock@openss7.org,
        qxie1@email.mot.com, s_balani@trillium.com, Qiaobing.Xie@motorola.com,
        mo_khan@trillium.com, h_pillai@trillium.com, jheitz@lucent.com,
        tsvwg@ietf.org
To: Jon Grimm <jgrimm@austin.ibm.com>
From: Michael Tuexen <Michael.Tuexen@icn.siemens.de>
In-Reply-To: <3CCDA2AD.CB11491@austin.ibm.com>
Message-Id: <3A3E82F9-5BAD-11D6-ACAA-0030654C1AB6@icn.siemens.de>
Content-Transfer-Encoding: 7bit
X-Mailer: Apple Mail (2.481)
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Dear all,

one possibility to capture this stuff is in an separate ID. It could
deal with such terminology stuff and usage stuff for AddIP. Maybe
an applicability statement or BCP. Don't know, but it could life for
some time as an ID and become an RFC later on.
If this seems a good idea I could try to write up some text ...

Best regards
Michael

On Monday, April 29, 2002, at 09:44 PM, Jon Grimm wrote:

> Thanks for the remarks, Randy.
>  I have a few comments still below.  Thanks, Jon
>
> Randall Stewart wrote:
>>
>> Jon:
>>
>> a couple of comments :>
>>
>> Jon Grimm wrote:
>>>
>>> Ivan/All
>>>
>>>         I think something that needs to come out of this discussion 
>>> are
>>> scenarios and mandatory behavior in those scenarios (or why 
>>> flexibility
>>> is allowed if not mandatory behavior).  Its a little disturbing that a
>>> similar 'what is an endpoint?' discussion seems to pop up again every 
>>> so
>>> often.  (e.g.bindx discussion last Sept in
>>> ftp://ftp.ietf.org/ietf-mail-archive/tsvwg/2001-09.mail ) and still 
>>> does
>>> not seem to be completely clear even to those following SCTP closely
>>> over the last year.
>>
>> I agree.. it is a discussion that seems to NEVER end :>
>>
>>>
>>> For each of the following scenarios,
>>>
>>> 1) A new address is to be added an endpoint. (manual)
>>> 2) An address is to be deleted from an endpoint.  (manual)
>>> 3) A new NIC/address is configured and an endpoint is bound to all
>>> (auto)
>>> 4) An address bound is brought down (no longer available) (auto).
>>> 5) An address bound is brought down for maintenance and then brought
>>> on-line again. (auto)
>>>
>>> consider each of the following subcases (assuming multiple 
>>> associations
>>> under an ep).
>>>
>>> a) All Peers have ADD-IP.
>>> b) No Peers have ADD-IP.
>>> c) A subset of Peers support ADD-IP.
>>>
>>> What happens?
>>>
>>
>> Where are you proposing putting these examples? I think some
>> text is needed for clearification but I am not sure that a protocol
>> spec is where you illustrate all these different examples...
>
>
> In some I-D,of course.   I see you are considering the Sockets API I-D,
> yes this could work for me as an appendix there.  Given this latest
> round of discussion, folks may wildly disagree, but I'm coming from the
> angle that I just want it captured.  ;-)
>
>>
>>> Another comment:
>>> Dropping down a little bit into the Sockets API for illustration:
>>>
>>> Depending on one's interpretation of endpoint, what is the endpoint 
>>> for
>>> the association of a peeled-off UDP style socket or for that matter an
>>> 'accepted' TCP-style socket.  If the strict interpretation of the
>>> endpoint is kept.. wouldn't an ADD-IP on the child sockets (peeled off
>>> or accept TCP) affect the local endpoint of the parent socket?  Or 
>>> vise
>>> versa, should an ADD-IP on a TCP-style listening socket, affect the
>>> accepted sockets?  If yes, was this expected?  If no, does the 
>>> original
>>
>> In my view the ADD-IP is done against the listening type endpoint. So
>> in the case of connected and/or peeled off associations they are
>> a different critter. In some ways they have formed a "new" endpoint. 
>> But
>> a interesting thing about them is they CAN NOT listen. So maybe in
>> this case Brian's adding an address to the association is the right
>> approach. They can NEVER receive a new association.. only the single
>> listner (in the TCP or UDP case) can do that.
>>
>> So a peeled off or "connected" socket with only one association may
>> well be a candidate to do an add-ip on an association basis.. but not
>> on the listening endpoint or a UDP model grouping of associations...
>>
>
>
> Has this gone full circle then?   This sounds more like you are always
> doing add-ip on associations.   Or you are really always doing add-ip on
> (Brian's or Ivan's ??) individual communication endpoints that belong to
> a grouping (group of 1 (peeled off) or many (udp style))?
>
> Consider two cases:
> 1)  If not bound-all, and an address is added via bindx.. I'd want the
> bindx to apply to the listening endpoint, as well as the associations
> currently grouped under that socket, but not the peeled-off socket.
> 2) If however, if the original socket was bound-all and a new interface
> is configured, I'd want to this new information to propogate both to the
> original listening socket and any peeled off sockets.
>
> Do these make sense?
>
>
> A related question:  should Add-ip happen on associations waiting in the
> TCP-style accept queue?   manual bindx()?  auto INADDR_ANY??
>
>
>
>>> definition of endpoint still hold?
>>>
>>> Even if each expected behavior does not get spelled it, it would be 
>>> nice
>>> for the SCTP Sockets API I-D to clarify the relationship between a
>>> socket and an endpoint, so implementors can at least infer what the
>>> intended behavior should be.
>>
>> Hmm this may be the best place for this information since it will
>> be informational..
>
> Sure.  If thats were folks want it, as long as gets agreed on and
> captured somewhere.
>
> Thanks,
> Jon
>
>
>>
>> R
>>
>> --
>> Randall R. Stewart
>> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
>>
>> _______________________________________________
>> tsvwg mailing list
>> tsvwg@ietf.org
>> https://www1.ietf.org/mailman/listinfo/tsvwg
>>
Michael.Tuexen@micmac.franken.de



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 17:03:42 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA19882
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 17:03:42 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA22268
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 17:03:44 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA18537;
	Mon, 29 Apr 2002 16:17:30 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA17616
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 16:08:51 -0400 (EDT)
Received: from motgate4.mot.com (motgate4.mot.com [144.189.100.102])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA17826
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 16:08:48 -0400 (EDT)
Received: [from pobox2.mot.com (pobox2.mot.com [136.182.15.8]) by motgate4.mot.com (motgate4 2.1) with ESMTP id NAA26940; Mon, 29 Apr 2002 13:08:49 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox2.mot.com (MOT-pobox2 2.0) with ESMTP id NAA06682; Mon, 29 Apr 2002 13:08:49 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3TK8kb12891;
	Mon, 29 Apr 2002 15:08:46 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id PAA21828;
	Mon, 29 Apr 2002 15:08:56 -0500 (CDT)
Message-ID: <3CCDA8FF.4CAF72E1@email.mot.com>
Date: Mon, 29 Apr 2002 15:11:43 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Randall Stewart <randall@stewart.chicago.il.us>
CC: Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Tuexen Michael <Michael.Tuexen@icn.siemens.de>,
        Ivan.Arias-Rodriguez@nokia.com, s_balani@trillium.com,
        mo_khan@trillium.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F92978223B01D311ACFF0008C71EE19D015DF7D2@MCHH225E> <3CCD857A.7902D0D@email.mot.com> <3CCD93F3.CA06B1E@stewart.chicago.il.us>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

...
> I don't agree with this rule.. just because a peer in a UDP model
> endpoint on one of its assocaitions is NOT capable of add-ip why
> should the rest be stuck. Yes you have an abnormality here in
> that that peer COULD send a INIT to the new address.. again I repeat:
> 1) This is such a corner case it is almost silly.
> 2) If it happens it is not that hard to handle... Looking at
>    the code Peter and I wrote you would end up finding the endpoint
>    (since the address WAS added there) and you would then LOOK at
>    its associations. This would then appear to be a RESTART where the
>    peer added addresses. You would then SEND IT an ABORT saying restart
>    while adding addresses. This would leave the other association
>    un-touched and would elimiate the error on the peers part...

Randy,

Let's walk through this example in detail:
Say, local endpoint E1 has two associations to different peers,

     E1             A1     E2
[IPA,IPB:Port1] <-------> [...] 
     E1             A2     E3
[IPA,IPB:Port1] <-------> [IPZ:Port2] 

We do add-ip (IPC) to E1. After sending ASCONF to both E2 and E3, we
succeeded with E2 but failed on E3.

     E1'               A1'     E2
[IPA,IPB,IPC:Port1] <-------> [...] 
     E1'               A2'     E3
[IPA,IPB,IPC:Port1] <-------> [IPZ:Port2]  /* E3 rejected add-ip and */
                                      /* hence knows nothg about IPC */

My understanding is that you would like to allow this. 

Now, several things could happen,

1) E3 may get a packet from E1' with src addr=IPC. This packet will
becomes an OOTB to E3 and E3 will send back an ABORT to kill the
ill-fated A2'. This is ok.

2) E3 may send an INIT to IPC/Port1. Is this the corner case you
described above? If so, I can see E1'1 will find the association A2' and
take this E3 INIT as a restart, but I don't see why E1' will consider it
as a "restart with added address" (E3 didn't add any address, it still
uses only IPZ).

-Qiaobing


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 17:21:19 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA20353
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 17:21:19 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA22842
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 17:21:22 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA20993;
	Mon, 29 Apr 2002 16:58:22 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA20964
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 16:58:20 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA19626
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 16:58:15 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3TKvtH60376;
	Mon, 29 Apr 2002 15:57:56 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCDB3D3.B8C4964E@stewart.chicago.il.us>
Date: Mon, 29 Apr 2002 15:57:55 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Qiaobing Xie <qxie1@email.mot.com>
CC: Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Tuexen Michael <Michael.Tuexen@icn.siemens.de>,
        Ivan.Arias-Rodriguez@nokia.com, s_balani@trillium.com,
        mo_khan@trillium.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F92978223B01D311ACFF0008C71EE19D015DF7D2@MCHH225E> <3CCD857A.7902D0D@email.mot.com> <3CCD93F3.CA06B1E@stewart.chicago.il.us> <3CCDA8FF.4CAF72E1@email.mot.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Qiaobing Xie wrote:
> 
> ...
> > I don't agree with this rule.. just because a peer in a UDP model
> > endpoint on one of its assocaitions is NOT capable of add-ip why
> > should the rest be stuck. Yes you have an abnormality here in
> > that that peer COULD send a INIT to the new address.. again I repeat:
> > 1) This is such a corner case it is almost silly.
> > 2) If it happens it is not that hard to handle... Looking at
> >    the code Peter and I wrote you would end up finding the endpoint
> >    (since the address WAS added there) and you would then LOOK at
> >    its associations. This would then appear to be a RESTART where the
> >    peer added addresses. You would then SEND IT an ABORT saying restart
> >    while adding addresses. This would leave the other association
> >    un-touched and would elimiate the error on the peers part...
> 
> Randy,
> 
> Let's walk through this example in detail:
> Say, local endpoint E1 has two associations to different peers,
> 
>      E1             A1     E2
> [IPA,IPB:Port1] <-------> [...]
>      E1             A2     E3
> [IPA,IPB:Port1] <-------> [IPZ:Port2]
> 
> We do add-ip (IPC) to E1. After sending ASCONF to both E2 and E3, we
> succeeded with E2 but failed on E3.
> 
>      E1'               A1'     E2
> [IPA,IPB,IPC:Port1] <-------> [...]
>      E1'               A2'     E3
> [IPA,IPB,IPC:Port1] <-------> [IPZ:Port2]  /* E3 rejected add-ip and */
>                                       /* hence knows nothg about IPC */
> 
> My understanding is that you would like to allow this.
> 
> Now, several things could happen,
> 
> 1) E3 may get a packet from E1' with src addr=IPC. This packet will
> becomes an OOTB to E3 and E3 will send back an ABORT to kill the
> ill-fated A2'. This is ok.
> 

Why would E1' use IPC... in this case it knows that it is NOT
part of the association. So it MUST NEVER source things from
IPC as long as this holds true .. no problem.. only if
E1's implementation is so bad it sources from IPC that a
problem will occur.


> 2) E3 may send an INIT to IPC/Port1. Is this the corner case you
> described above? If so, I can see E1'1 will find the association A2' and
> take this E3 INIT as a restart, but I don't see why E1' will consider it
> as a "restart with added address" (E3 didn't add any address, it still
> uses only IPZ).

Yes.. this is a very corner case.. 

 A) E3 must be very confused to want two associations.. so in
    my mind this is a corner case.

 B) We would need to expand the rules that you must also check
    the your addresses in a COOKIE-ECHO to see if you have
    hit such a collision.. if so send an ABORT... 

I don't see this is a problem.. ABORTING your funny association
with E3 should leave A2 alone and all should be well.

It seems to me we need to add this new rule for ADD-IP.

R

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 17:46:44 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA21030
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 17:46:44 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA24270
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 17:46:47 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA22820;
	Mon, 29 Apr 2002 17:20:02 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA22791
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 17:20:00 -0400 (EDT)
Received: from caduceus.fm.intel.com (fmr02.intel.com [192.55.52.25])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA20327
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 17:19:56 -0400 (EDT)
Received: from talaria.fm.intel.com (talaria.fm.intel.com [10.1.192.39])
	by caduceus.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.42 2002/04/26 23:25:13 root Exp $) with ESMTP id g3TLJ0p09248
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 21:19:05 GMT
Received: from fmsmsxvs041.fm.intel.com (fmsmsxv041-1.fm.intel.com [132.233.48.109])
	by talaria.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.17 2002/04/27 00:24:04 root Exp $) with SMTP id g3TLMCi26656
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 21:22:12 GMT
Received: from fmsmsx29.FM.INTEL.COM ([132.233.42.29])
 by fmsmsxvs041.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002042914201205581
 ; Mon, 29 Apr 2002 14:20:12 -0700
Received: by fmsmsx29.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <J458WN31>; Mon, 29 Apr 2002 14:19:52 -0700
Message-ID: <F1CE15E08172D4119247009027AE9D50191173AD@FMSMSX37>
From: "Balani, Sandeep" <s_balani@trillium.com>
To: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        Qiaobing Xie <qxie1@email.mot.com>
Cc: Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Tuexen Michael<Michael.Tuexen@icn.siemens.de>,
        Ivan.Arias-Rodriguez@nokia.com,
        "Balani, Sandeep" <s_balani@trillium.com>,
        "Khan, Mohammad"<mo_khan@trillium.com>, tsvwg@ietf.org
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Mon, 29 Apr 2002 14:19:44 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

hi all, 
  Please see my responses below: 

> -----Original Message-----
> From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: Monday, April 29, 2002 1:58 PM
> To: Qiaobing Xie
> Cc: Qiaobing Xie; Tuexen Michael; Ivan.Arias-Rodriguez@nokia.com;
> s_balani@trillium.com; mo_khan@trillium.com; tsvwg@ietf.org
> Subject: Re: [Tsvwg] End point and association Address conflict
> 
> 
> Qiaobing Xie wrote:
> > 
> > ...
> > > I don't agree with this rule.. just because a peer in a UDP model
> > > endpoint on one of its assocaitions is NOT capable of add-ip why
> > > should the rest be stuck. Yes you have an abnormality here in
> > > that that peer COULD send a INIT to the new address.. 
> again I repeat:
> > > 1) This is such a corner case it is almost silly.
> > > 2) If it happens it is not that hard to handle... Looking at
> > >    the code Peter and I wrote you would end up finding 
> the endpoint
> > >    (since the address WAS added there) and you would then LOOK at
> > >    its associations. This would then appear to be a 
> RESTART where the
> > >    peer added addresses. You would then SEND IT an ABORT 
> saying restart
> > >    while adding addresses. This would leave the other association
> > >    un-touched and would elimiate the error on the peers part...
> > 
> > Randy,
> > 
> > Let's walk through this example in detail:
> > Say, local endpoint E1 has two associations to different peers,
> > 
> >      E1             A1     E2
> > [IPA,IPB:Port1] <-------> [...]
> >      E1             A2     E3
> > [IPA,IPB:Port1] <-------> [IPZ:Port2]
> > 
> > We do add-ip (IPC) to E1. After sending ASCONF to both E2 and E3, we
> > succeeded with E2 but failed on E3.
> > 
> >      E1'               A1'     E2
> > [IPA,IPB,IPC:Port1] <-------> [...]
> >      E1'               A2'     E3
> > [IPA,IPB,IPC:Port1] <-------> [IPZ:Port2]  /* E3 rejected 
> add-ip and */
> >                                       /* hence knows nothg 
> about IPC */
> > 
> > My understanding is that you would like to allow this.
> > 
> > Now, several things could happen,
> > 
> > 1) E3 may get a packet from E1' with src addr=IPC. This packet will
> > becomes an OOTB to E3 and E3 will send back an ABORT to kill the
> > ill-fated A2'. This is ok.
> > 
> 
> Why would E1' use IPC... in this case it knows that it is NOT
> part of the association. So it MUST NEVER source things from
> IPC as long as this holds true .. no problem.. only if
> E1's implementation is so bad it sources from IPC that a
> problem will occur.
> 


[ Sandeep ] Hi Randy,  It is possible because some implementation don't
control source IP address the same example u used for INADDR_ANY. In that
case if one don't control source IP address A2 might receive packet from
IPc. 


> 
> > 2) E3 may send an INIT to IPC/Port1. Is this the corner case you
> > described above? If so, I can see E1'1 will find the 
> association A2' and
> > take this E3 INIT as a restart, but I don't see why E1' 
> will consider it
> > as a "restart with added address" (E3 didn't add any 
> address, it still
> > uses only IPZ).
> 
> Yes.. this is a very corner case.. 
> 
>  A) E3 must be very confused to want two associations.. so in
>     my mind this is a corner case.
> 
>  B) We would need to expand the rules that you must also check
>     the your addresses in a COOKIE-ECHO to see if you have
>     hit such a collision.. if so send an ABORT... 
> 
> I don't see this is a problem.. ABORTING your funny association
> with E3 should leave A2 alone and all should be well.
> 
> It seems to me we need to add this new rule for ADD-IP.
> 
> R
> 


 [ Sandeep ] Actually this type of rule bring out various possiblities.
Please consider below : 


Let assume the same example given by Q : 

      E1             A1     E2
 [IPA,IPB:Port1] <-------> [...]
     E1             A2     E3
 [IPA,IPB:Port1] <-------> [IPZ:Port2]
 
 We do add-ip (IPC) to E1. After sending ASCONF to both E2 and E3, we
 succeeded with E2 but failed on E3.

     E1'               A1'     E2
[IPA,IPB,IPC:Port1] <-------> [...]
     E1'               A2'     E3
[IPA,IPB,IPC:Port1] <-------> [IPZ:Port2]  /* E3 rejected add-ip and */


Now as we are adding the IP addresses on Endpoint basis there are more
collision cases in these sceinarios : 

Scenario 1) 
============== 
1) U  have an Endpoint E1 [ IPa, IPb, Port1 ] with Assocations A1, & A2. 
2) Then u receive another INIT chunk for endpoint E1, U send INIT-ACK[ with
IPa, IPb ] to for Assoc A3. 
3) In the mean time u receive COOKIE-ECHO back from A3, lets say we receive
ADDIP for Endpoint E1 to add another IP address IPc, Now We add IPc in E1
and send ASCONF chunk to A1. 
4) After this step lets say we receive COOKIE-ECHO with from A3 and in worst
case our association will be established with subset of address IPa, IPb,
which was not intended. 

Sceinario 2) 
=============
1) U  have an Endpoint E1 [ IPa, IPb, Port1 ] with Assocations A1, & A2. 
2) Then u receive another INIT chunk for endpoint E1, U send INIT-ACK[ with
IPa, IPb ] to for Assoc A3. 
3) In the mean time u receive COOKIE-ECHO back from A3, lets say we receive
DELETEIP for Endpoint E1 to delete address IPa, Now We delete IPa from E1
and send ASCONF chunk to A1. 
4) After this  lets say we receive COOKIE-ECHO with from A3 and if i apply
Randall's rule i will be sending ABORT for this association A3, Which also
was not intended. 


umm ... there may be some more cases left ...... But is'nt there any
simple/clean solution for all this  ??????

thanks
sandeep 


> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 18:44:47 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA22003
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 18:44:42 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id SAA27618
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 18:44:46 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA25481;
	Mon, 29 Apr 2002 18:16:07 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA25399
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 18:16:01 -0400 (EDT)
Received: from caduceus.fm.intel.com (fmr02.intel.com [192.55.52.25])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA21473
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 18:15:54 -0400 (EDT)
Received: from petasus.fm.intel.com (petasus.fm.intel.com [10.1.192.37])
	by caduceus.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.42 2002/04/26 23:25:13 root Exp $) with ESMTP id g3TMExp17625
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 22:15:00 GMT
Received: from fmsmsxvs042.fm.intel.com (fmsmsxv042-1.fm.intel.com [132.233.48.110])
	by petasus.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.17 2002/04/27 00:24:04 root Exp $) with SMTP id g3TMEDw19772
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 22:14:13 GMT
Received: from FMSMSX018.fm.intel.com ([132.233.42.197])
 by fmsmsxvs042.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002042915200925783
 ; Mon, 29 Apr 2002 15:20:09 -0700
Received: by fmsmsx018.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <2WWTYHY5>; Mon, 29 Apr 2002 15:15:50 -0700
Message-ID: <F1CE15E08172D4119247009027AE9D50191173AE@FMSMSX37>
From: "Balani, Sandeep" <s_balani@trillium.com>
To: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        Qiaobing Xie <qxie1@email.mot.com>
Cc: Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Tuexen Michael<Michael.Tuexen@icn.siemens.de>,
        Ivan.Arias-Rodriguez@nokia.com,
        "Balani, Sandeep" <s_balani@trillium.com>,
        "Khan, Mohammad"<mo_khan@trillium.com>, tsvwg@ietf.org
Subject: RE: [Tsvwg] End point and association Address conflict
Date: Mon, 29 Apr 2002 15:15:40 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi all,  

   I was looking at AddIP draft in which Section 4.1.1 has following RULE :

    R1) One and only one ASCONF Chunk MAY be in transit and
    unacknowledged at any one time.  If a sender, after sending an ASCONF
    chunk, decides it needs to transfer another ASCONF Chunk, it MUST
    wait until the ASCONF-ACK Chunk returns from the previous ASCONF
    Chunk before sending a subsequent ASCONF. Note this restriction
    binds each side, so at any time two ASCONF may be in-transit on any
    given association (one sent from each endpoint).


If thats the case, And let say we have an M3UA Endpoint at SGP node as E1 [
IP-A, 2905 ] & this SGP has associations with 100 ASP's .   And in this case
if i do ADDIP(IPb) on this Endpoint E1, than I have a major problem because
i have to send 100 ASCONF chunks one-by-one on all associations, which seems
very difficult & cumbersome taking in to consideration the race conditions
and scenarios mentioned in my previous mails. 


Any comments ???

thanks
sandeep  



> -----Original Message-----
> From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: Monday, April 29, 2002 1:58 PM
> To: Qiaobing Xie
> Cc: Qiaobing Xie; Tuexen Michael; Ivan.Arias-Rodriguez@nokia.com;
> s_balani@trillium.com; mo_khan@trillium.com; tsvwg@ietf.org
> Subject: Re: [Tsvwg] End point and association Address conflict
> 
> 
> Qiaobing Xie wrote:
> > 
> > ...
> > > I don't agree with this rule.. just because a peer in a UDP model
> > > endpoint on one of its assocaitions is NOT capable of add-ip why
> > > should the rest be stuck. Yes you have an abnormality here in
> > > that that peer COULD send a INIT to the new address.. 
> again I repeat:
> > > 1) This is such a corner case it is almost silly.
> > > 2) If it happens it is not that hard to handle... Looking at
> > >    the code Peter and I wrote you would end up finding 
> the endpoint
> > >    (since the address WAS added there) and you would then LOOK at
> > >    its associations. This would then appear to be a 
> RESTART where the
> > >    peer added addresses. You would then SEND IT an ABORT 
> saying restart
> > >    while adding addresses. This would leave the other association
> > >    un-touched and would elimiate the error on the peers part...
> > 
> > Randy,
> > 
> > Let's walk through this example in detail:
> > Say, local endpoint E1 has two associations to different peers,
> > 
> >      E1             A1     E2
> > [IPA,IPB:Port1] <-------> [...]
> >      E1             A2     E3
> > [IPA,IPB:Port1] <-------> [IPZ:Port2]
> > 
> > We do add-ip (IPC) to E1. After sending ASCONF to both E2 and E3, we
> > succeeded with E2 but failed on E3.
> > 
> >      E1'               A1'     E2
> > [IPA,IPB,IPC:Port1] <-------> [...]
> >      E1'               A2'     E3
> > [IPA,IPB,IPC:Port1] <-------> [IPZ:Port2]  /* E3 rejected 
> add-ip and */
> >                                       /* hence knows nothg 
> about IPC */
> > 
> > My understanding is that you would like to allow this.
> > 
> > Now, several things could happen,
> > 
> > 1) E3 may get a packet from E1' with src addr=IPC. This packet will
> > becomes an OOTB to E3 and E3 will send back an ABORT to kill the
> > ill-fated A2'. This is ok.
> > 
> 
> Why would E1' use IPC... in this case it knows that it is NOT
> part of the association. So it MUST NEVER source things from
> IPC as long as this holds true .. no problem.. only if
> E1's implementation is so bad it sources from IPC that a
> problem will occur.
> 
> 
> > 2) E3 may send an INIT to IPC/Port1. Is this the corner case you
> > described above? If so, I can see E1'1 will find the 
> association A2' and
> > take this E3 INIT as a restart, but I don't see why E1' 
> will consider it
> > as a "restart with added address" (E3 didn't add any 
> address, it still
> > uses only IPZ).
> 
> Yes.. this is a very corner case.. 
> 
>  A) E3 must be very confused to want two associations.. so in
>     my mind this is a corner case.
> 
>  B) We would need to expand the rules that you must also check
>     the your addresses in a COOKIE-ECHO to see if you have
>     hit such a collision.. if so send an ABORT... 
> 
> I don't see this is a problem.. ABORTING your funny association
> with E3 should leave A2 alone and all should be well.
> 
> It seems to me we need to add this new rule for ADD-IP.
> 
> R
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 18:58:22 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA22184
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 18:58:21 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id SAA28098
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 18:58:23 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA27034;
	Mon, 29 Apr 2002 18:32:36 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA27003
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 18:32:34 -0400 (EDT)
Received: from rwcrmhc53.attbi.com (rwcrmhc53.attbi.com [204.127.198.39])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA21740
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 18:32:29 -0400 (EDT)
Received: from ieee.org ([12.248.77.94]) by rwcrmhc53.attbi.com
          (InterMail vM.4.01.03.27 201-229-121-127-20010626) with ESMTP
          id <20020429223202.CHSD1392.rwcrmhc53.attbi.com@ieee.org>;
          Mon, 29 Apr 2002 22:32:02 +0000
Message-ID: <3CCDC94A.F2745C47@ieee.org>
Date: Mon, 29 Apr 2002 17:29:30 -0500
From: Peter Lei <peter.lei@ieee.org>
X-Mailer: Mozilla 4.79 [en] (Windows NT 5.0; U)
X-Accept-Language: en
MIME-Version: 1.0
To: "Balani, Sandeep" <s_balani@trillium.com>
CC: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        Qiaobing Xie <qxie1@email.mot.com>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Tuexen Michael <Michael.Tuexen@icn.siemens.de>,
        Ivan.Arias-Rodriguez@nokia.com,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F1CE15E08172D4119247009027AE9D50191173AE@FMSMSX37>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

> Hi all,
> 
>    I was looking at AddIP draft in which Section 4.1.1 has following RULE :
> 
>     R1) One and only one ASCONF Chunk MAY be in transit and
>     unacknowledged at any one time.  If a sender, after sending an ASCONF
>     chunk, decides it needs to transfer another ASCONF Chunk, it MUST
>     wait until the ASCONF-ACK Chunk returns from the previous ASCONF
>     Chunk before sending a subsequent ASCONF. Note this restriction
>     binds each side, so at any time two ASCONF may be in-transit on any
>     given association (one sent from each endpoint).
> 
> If thats the case, And let say we have an M3UA Endpoint at SGP node as E1 [
> IP-A, 2905 ] & this SGP has associations with 100 ASP's .   And in this case
> if i do ADDIP(IPb) on this Endpoint E1, than I have a major problem because
> i have to send 100 ASCONF chunks one-by-one on all associations, which seems
> very difficult & cumbersome taking in to consideration the race conditions
> and scenarios mentioned in my previous mails.
>
> Any comments ???


That should be one asconf outstanding per association; see the last sentence,
only two asconf's can be in-flight "on any given association".

--peter

> 
> thanks
> sandeep
> 
> > -----Original Message-----
> > From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> > Sent: Monday, April 29, 2002 1:58 PM
> > To: Qiaobing Xie
> > Cc: Qiaobing Xie; Tuexen Michael; Ivan.Arias-Rodriguez@nokia.com;
> > s_balani@trillium.com; mo_khan@trillium.com; tsvwg@ietf.org
> > Subject: Re: [Tsvwg] End point and association Address conflict
> >
> >
> > Qiaobing Xie wrote:
> > >
> > > ...
> > > > I don't agree with this rule.. just because a peer in a UDP model
> > > > endpoint on one of its assocaitions is NOT capable of add-ip why
> > > > should the rest be stuck. Yes you have an abnormality here in
> > > > that that peer COULD send a INIT to the new address..
> > again I repeat:
> > > > 1) This is such a corner case it is almost silly.
> > > > 2) If it happens it is not that hard to handle... Looking at
> > > >    the code Peter and I wrote you would end up finding
> > the endpoint
> > > >    (since the address WAS added there) and you would then LOOK at
> > > >    its associations. This would then appear to be a
> > RESTART where the
> > > >    peer added addresses. You would then SEND IT an ABORT
> > saying restart
> > > >    while adding addresses. This would leave the other association
> > > >    un-touched and would elimiate the error on the peers part...
> > >
> > > Randy,
> > >
> > > Let's walk through this example in detail:
> > > Say, local endpoint E1 has two associations to different peers,
> > >
> > >      E1             A1     E2
> > > [IPA,IPB:Port1] <-------> [...]
> > >      E1             A2     E3
> > > [IPA,IPB:Port1] <-------> [IPZ:Port2]
> > >
> > > We do add-ip (IPC) to E1. After sending ASCONF to both E2 and E3, we
> > > succeeded with E2 but failed on E3.
> > >
> > >      E1'               A1'     E2
> > > [IPA,IPB,IPC:Port1] <-------> [...]
> > >      E1'               A2'     E3
> > > [IPA,IPB,IPC:Port1] <-------> [IPZ:Port2]  /* E3 rejected
> > add-ip and */
> > >                                       /* hence knows nothg
> > about IPC */
> > >
> > > My understanding is that you would like to allow this.
> > >
> > > Now, several things could happen,
> > >
> > > 1) E3 may get a packet from E1' with src addr=IPC. This packet will
> > > becomes an OOTB to E3 and E3 will send back an ABORT to kill the
> > > ill-fated A2'. This is ok.
> > >
> >
> > Why would E1' use IPC... in this case it knows that it is NOT
> > part of the association. So it MUST NEVER source things from
> > IPC as long as this holds true .. no problem.. only if
> > E1's implementation is so bad it sources from IPC that a
> > problem will occur.
> >
> >
> > > 2) E3 may send an INIT to IPC/Port1. Is this the corner case you
> > > described above? If so, I can see E1'1 will find the
> > association A2' and
> > > take this E3 INIT as a restart, but I don't see why E1'
> > will consider it
> > > as a "restart with added address" (E3 didn't add any
> > address, it still
> > > uses only IPZ).
> >
> > Yes.. this is a very corner case..
> >
> >  A) E3 must be very confused to want two associations.. so in
> >     my mind this is a corner case.
> >
> >  B) We would need to expand the rules that you must also check
> >     the your addresses in a COOKIE-ECHO to see if you have
> >     hit such a collision.. if so send an ABORT...
> >
> > I don't see this is a problem.. ABORTING your funny association
> > with E3 should leave A2 alone and all should be well.
> >
> > It seems to me we need to add this new rule for ADD-IP.
> >
> > R
> >
> > --
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> >
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 19:34:19 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA23178
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 19:34:19 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id TAA02337
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 19:34:21 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA29086;
	Mon, 29 Apr 2002 19:06:04 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA29061
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 19:06:02 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA22402
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 19:05:59 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3TN5eH60596;
	Mon, 29 Apr 2002 18:05:40 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCDD1C3.B9F97246@stewart.chicago.il.us>
Date: Mon, 29 Apr 2002 18:05:40 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: "Balani, Sandeep" <s_balani@trillium.com>
CC: Qiaobing Xie <qxie1@email.mot.com>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Tuexen Michael <Michael.Tuexen@icn.siemens.de>,
        Ivan.Arias-Rodriguez@nokia.com,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F1CE15E08172D4119247009027AE9D50191173AD@FMSMSX37>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Sandeep:

comments below

"Balani, Sandeep" wrote:
> 
> hi all,
>   Please see my responses below:
> 
> > -----Original Message-----
> > From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> > Sent: Monday, April 29, 2002 1:58 PM
> > To: Qiaobing Xie
> > Cc: Qiaobing Xie; Tuexen Michael; Ivan.Arias-Rodriguez@nokia.com;
> > s_balani@trillium.com; mo_khan@trillium.com; tsvwg@ietf.org
> > Subject: Re: [Tsvwg] End point and association Address conflict
> >
> >
> > Qiaobing Xie wrote:
> > >
> > > ...
> > > > I don't agree with this rule.. just because a peer in a UDP model
> > > > endpoint on one of its assocaitions is NOT capable of add-ip why
> > > > should the rest be stuck. Yes you have an abnormality here in
> > > > that that peer COULD send a INIT to the new address..
> > again I repeat:
> > > > 1) This is such a corner case it is almost silly.
> > > > 2) If it happens it is not that hard to handle... Looking at
> > > >    the code Peter and I wrote you would end up finding
> > the endpoint
> > > >    (since the address WAS added there) and you would then LOOK at
> > > >    its associations. This would then appear to be a
> > RESTART where the
> > > >    peer added addresses. You would then SEND IT an ABORT
> > saying restart
> > > >    while adding addresses. This would leave the other association
> > > >    un-touched and would elimiate the error on the peers part...
> > >
> > > Randy,
> > >
> > > Let's walk through this example in detail:
> > > Say, local endpoint E1 has two associations to different peers,
> > >
> > >      E1             A1     E2
> > > [IPA,IPB:Port1] <-------> [...]
> > >      E1             A2     E3
> > > [IPA,IPB:Port1] <-------> [IPZ:Port2]
> > >
> > > We do add-ip (IPC) to E1. After sending ASCONF to both E2 and E3, we
> > > succeeded with E2 but failed on E3.
> > >
> > >      E1'               A1'     E2
> > > [IPA,IPB,IPC:Port1] <-------> [...]
> > >      E1'               A2'     E3
> > > [IPA,IPB,IPC:Port1] <-------> [IPZ:Port2]  /* E3 rejected
> > add-ip and */
> > >                                       /* hence knows nothg
> > about IPC */
> > >
> > > My understanding is that you would like to allow this.
> > >
> > > Now, several things could happen,
> > >
> > > 1) E3 may get a packet from E1' with src addr=IPC. This packet will
> > > becomes an OOTB to E3 and E3 will send back an ABORT to kill the
> > > ill-fated A2'. This is ok.
> > >
> >
> > Why would E1' use IPC... in this case it knows that it is NOT
> > part of the association. So it MUST NEVER source things from
> > IPC as long as this holds true .. no problem.. only if
> > E1's implementation is so bad it sources from IPC that a
> > problem will occur.
> >
> 
> [ Sandeep ] Hi Randy,  It is possible because some implementation don't
> control source IP address the same example u used for INADDR_ANY. In that
> case if one don't control source IP address A2 might receive packet from
> IPc.
> 
> >
> > > 2) E3 may send an INIT to IPC/Port1. Is this the corner case you
> > > described above? If so, I can see E1'1 will find the
> > association A2' and
> > > take this E3 INIT as a restart, but I don't see why E1'
> > will consider it
> > > as a "restart with added address" (E3 didn't add any
> > address, it still
> > > uses only IPZ).
> >
> > Yes.. this is a very corner case..
> >
> >  A) E3 must be very confused to want two associations.. so in
> >     my mind this is a corner case.
> >
> >  B) We would need to expand the rules that you must also check
> >     the your addresses in a COOKIE-ECHO to see if you have
> >     hit such a collision.. if so send an ABORT...
> >
> > I don't see this is a problem.. ABORTING your funny association
> > with E3 should leave A2 alone and all should be well.
> >
> > It seems to me we need to add this new rule for ADD-IP.
> >
> > R
> >
> 
>  [ Sandeep ] Actually this type of rule bring out various possiblities.
> Please consider below :
> 
> Let assume the same example given by Q :
> 
>       E1             A1     E2
>  [IPA,IPB:Port1] <-------> [...]
>      E1             A2     E3
>  [IPA,IPB:Port1] <-------> [IPZ:Port2]
> 
>  We do add-ip (IPC) to E1. After sending ASCONF to both E2 and E3, we
>  succeeded with E2 but failed on E3.
> 
>      E1'               A1'     E2
> [IPA,IPB,IPC:Port1] <-------> [...]
>      E1'               A2'     E3
> [IPA,IPB,IPC:Port1] <-------> [IPZ:Port2]  /* E3 rejected add-ip and */
> 
> Now as we are adding the IP addresses on Endpoint basis there are more
> collision cases in these sceinarios :
> 
> Scenario 1)
> ==============
> 1) U  have an Endpoint E1 [ IPa, IPb, Port1 ] with Assocations A1, & A2.
> 2) Then u receive another INIT chunk for endpoint E1, U send INIT-ACK[ with
> IPa, IPb ] to for Assoc A3.

Why would you do this? You just added IPc to your E1.. you would
send an INIT-ACK with [IPa,IPb and IPc]... however if you are
saying it is inbetween.. i.e. you send the INIT-ACK and
THEN IPc is added.. yes that can happen and in such a case
I believe you must look at the cookie and recognize you must
do a ADD-IP on the IPc.. I think you will find this is what our
BSD code does :>


> 3) In the mean time u receive COOKIE-ECHO back from A3, lets say we receive
> ADDIP for Endpoint E1 to add another IP address IPc, Now We add IPc in E1
> and send ASCONF chunk to A1.

Ok.. 

and then you would also queue up the ASCONF for the A2 and A3..

> 4) After this step lets say we receive COOKIE-ECHO with from A3 and in worst
> case our association will be established with subset of address IPa, IPb,
> which was not intended.
> 

But you MUST look at the cookie echo and when you unpack it figure
out that you are missing an address.. I believe Peter put in
a whole module that handles this very race condition.. its part
of coding ... you have to try to cover the corner cases :>>


> Sceinario 2)
> =============
> 1) U  have an Endpoint E1 [ IPa, IPb, Port1 ] with Assocations A1, & A2.
> 2) Then u receive another INIT chunk for endpoint E1, U send INIT-ACK[ with
> IPa, IPb ] to for Assoc A3.
> 3) In the mean time u receive COOKIE-ECHO back from A3, lets say we receive
> DELETEIP for Endpoint E1 to delete address IPa, Now We delete IPa from E1
> and send ASCONF chunk to A1.
> 4) After this  lets say we receive COOKIE-ECHO with from A3 and if i apply
> Randall's rule i will be sending ABORT for this association A3, Which also
> was not intended.
> 

I am not convinced now we need a rule for it.. I will follow up
in another email... 

Now back to the example (forget my half-baked idea.. I have baked
it some more and don't like it :>) above you would realize (at unpack)
that you need to send a DELETE-IP.


> umm ... there may be some more cases left ...... But is'nt there any
> simple/clean solution for all this  ??????
> 

I will follow up .. after reading some more of the queued emails :-)

There is a big piece we have forgotten in this discussion that makes
it so we DO NOT need a rule like I earlier proposed in my 1/2 baked
thoughts...

R

> 
> > --
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> >

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 20:39:47 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA24393
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 20:39:47 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA06528
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 20:39:49 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA05298;
	Mon, 29 Apr 2002 20:16:49 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA28613
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 19:01:14 -0400 (EDT)
Received: from motgate4.mot.com (motgate4.mot.com [144.189.100.102])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA22286
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 19:01:11 -0400 (EDT)
Received: [from pobox3.mot.com (pobox3.mot.com [10.64.251.242]) by motgate4.mot.com (motgate4 2.1) with ESMTP id QAA05237; Mon, 29 Apr 2002 16:01:12 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox3.mot.com (MOT-pobox3 2.0) with ESMTP id PAA12561; Mon, 29 Apr 2002 15:48:29 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3TN19b26909;
	Mon, 29 Apr 2002 18:01:09 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id SAA21969;
	Mon, 29 Apr 2002 18:01:19 -0500 (CDT)
Message-ID: <3CCDD167.8A0DBC43@email.mot.com>
Date: Mon, 29 Apr 2002 18:04:07 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Randall Stewart <randall@stewart.chicago.il.us>
CC: Jon Grimm <jgrimm@austin.ibm.com>, Ivan.Arias-Rodriguez@nokia.com,
        bidulock@openss7.org, Michael.Tuexen@icn.siemens.de,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>, s_balani@trillium.com,
        mo_khan@trillium.com, h_pillai@trillium.com, jheitz@lucent.com,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFB4@esebe017.NOE.Nokia.com> <3CCD844A.DB1BC960@austin.ibm.com> <3CCD9267.6456A90E@stewart.chicago.il.us> <3CCDA2AD.CB11491@austin.ibm.com> <3CCDB210.E8FA4663@stewart.chicago.il.us>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi, Randy,

I think we are still using slightly different endpoint definitions (as
you also realized below).

...
> 
> I still view it as adding to an endpoint but I think we must
> re-think about this... I put this down in a private email discussion
> but I think I will put it out here now ... even though the
> thoughts are only partially baked :>
> 
> We have basically 4 types of endpoints/sockets (note I think of
> an endpoint as a socket aka open FD).

I think of an endpoint as an abstract logical entity (as dryly defined
in RFC2960). I see three levels of address sets involved here:

Level 0: host addr set "S" = all the network addresses assigned to a
host (host is logically represented by a single OS instance). A host can
harbor multiple sockets/open FDs.

Level 1: addr set bound to a socket/open FD "s" = a subset of S bound to
the FD (this is the logical representation of an SCTP stack). An open FD
can harbor multiple endpoints. But some implementation, for simplicity
reason, may support only one endpoint.

Level 2: addr set of an endpoint "e" = yet a subset of "s". This is the
set of addresses actually used (or to be used) in an association.

In summary, e <= s <= S.

All my previous discussions were based on the "e" definition (including
the rules I proposed). 
Therefore, the following statement from Ivan 

  - An association, can use a subset of the source addresses of the
    endpoint to which it belongs.

becomes correct if Ivan's "endpoint" here implies a socket/open FD, but
incorrect if the "endpoint" here is defined as "e" above.

I will think more about the socket implications.

-Qiaobing

> 
> Now we have:
> 
> 1) A UDP model socket/listener that A) accepts new associations and
>    has some number of associations under it (fanned out is the term
>    I think Qiaobing used).
> 
> 2) A TCP model listener. This guy will have NO associations under
>    it accept maybe onces waiting to be accepted ... nothing can
>    happen on those though.. until AFTER they are accepted.
> 
> 3) Accepted and or peel'ed off ones.. these are a single association.
> 
> 4) A TCP model connect() caller. This guy can NOT accept any.. but is
>    going to have a single "association" to a peer... note this may
>    be (in the end) to a <1> or <3> type guy.. it does not matter.
> 
> Now lets look at each case..
> 
> First, an easy one, <3> is something that can NEVER get any new
> associations. INIT's don't go to it... they go to whoever was the
> listner... so in this case this association COULD have add-ip go
> against it. .thoughts/notes...{I believe in our implementation of add-ip
> this
>                                guy MAY auto-add IP if it is setup via an
>                                sockopt to do it.. default may be on
> especially
>                                if the parent (listener) is bound all
>                                }
> 
> Now lets look at another easy one.. <4> the connector.. this guy
> IS NOT LISTENING.. it TOO can't get a INIT in.. an ABORT() would be
> sent back... At the termination of its association, however, it MAY
> do a listen() call..  so here we would ADD the address to the endpoint
> (if it was turned on socketopt wise to do so..) and the attempt if
> possible to add it to the peer.
> 
> Now this leaves case <1> and <2>. The only real difference between
> these guys is that we have one that has a cluster of associations
> tied to it.. the other does NOT. Now we here ADD the address to the
> endpoint. For <2> we are done.. when the next INIT comes in the new
> association will form... and we are ok... For <1> new associations
> will use the new address.. (all of this assumes that the option
> to add automatically is enabled.. for bind-all this is probably
> the default.. it may not be for sub-set bound). Now for the
> existing "fan" of associations we would individually attempt to
> add the address to them..
> 
> Now for the big question .. what happens if we get our silly endpoint
> (peer) that sends an INIT while holding an association to us (either
> a TCP one or a fanned out one). Well some observations:
> 
> a) First you must be able to find the old associations, if you
>    can't then your INIT collision and RESTART code is NOT working.
> 
> b) When you find the "old" associations it is going to look like
>    a restart condition... especially for <1> since you locate the
>    endpoint (and you have added the address here hopefully) and then
>    you look for the association via one of the peers addresses. For the
>    TCP model you also must switch the socket you are associated with
>    in this hunt... hmmm lets look at both sides..
>     - If you find it then you would actually go through the
>       restart scenario as long as the peer used the same addresses
>       it would appear to be a restart.. this would be very interesting
>       and of course the peer would get very messed up I bet... but its
>       a silly endpoint anyway.. in the end the old association would
>       die, one way or another... Now if we add a restriction that we
>       can't add any new addresses on either side (aka like our current
>       peer can't add new addresses) then no problem.. we would ABORT()
>       with new addresses...
>     - If you can't find it then what would happen is when the COOKIE
>       arrives you could since you could look at all your addresses
>       and validate that a collision is not at hand... and then send
>       an abort...
> 
>    Pesonnely I prefer the no new addresses on either side during a
>    collision.. this solves the problem as long as you can locate your
>    associations.. which I think you should be able to..
> 
>  Well .. there are some of my 1/2 baked thoughts... need to mull on
>  them a bit more..
> 
> R
> 
> --
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 20:40:09 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA24410
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 20:40:09 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA06555
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 20:40:11 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA05191;
	Mon, 29 Apr 2002 20:16:33 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA25972
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 18:24:01 -0400 (EDT)
Received: from ftpbox.mot.com (ftpbox.mot.com [129.188.136.101])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA21568
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 18:23:56 -0400 (EDT)
Received: [from pobox2.mot.com (pobox2.mot.com [136.182.15.8]) by ftpbox.mot.com (ftpbox 2.1) with ESMTP id PAA15098; Mon, 29 Apr 2002 15:24:00 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox2.mot.com (MOT-pobox2 2.0) with ESMTP id PAA15271; Mon, 29 Apr 2002 15:23:59 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3TMNpb24331;
	Mon, 29 Apr 2002 17:23:51 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id RAA21927;
	Mon, 29 Apr 2002 17:24:01 -0500 (CDT)
Message-ID: <3CCDC8A9.39265443@email.mot.com>
Date: Mon, 29 Apr 2002 17:26:49 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: "Balani, Sandeep" <s_balani@trillium.com>
CC: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Tuexen Michael <Michael.Tuexen@icn.siemens.de>,
        Ivan.Arias-Rodriguez@nokia.com,
        "Khan, Mohammad" <mo_khan@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <F1CE15E08172D4119247009027AE9D50191173AE@FMSMSX37>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi, Sandeep,

This rule should be applied _per association_. So you are allowed one
ASCONF in-flight for each of the 100 fan-out associations in your
example. In other words, your local endpoint can fire out 100 ASCONFs,
each to a diffenrnt peer, at the same time. 

-Qiaobing

"Balani, Sandeep" wrote:
> 
> Hi all,
> 
>    I was looking at AddIP draft in which Section 4.1.1 has following RULE :
> 
>     R1) One and only one ASCONF Chunk MAY be in transit and
>     unacknowledged at any one time.  If a sender, after sending an ASCONF
>     chunk, decides it needs to transfer another ASCONF Chunk, it MUST
>     wait until the ASCONF-ACK Chunk returns from the previous ASCONF
>     Chunk before sending a subsequent ASCONF. Note this restriction
>     binds each side, so at any time two ASCONF may be in-transit on any
>     given association (one sent from each endpoint).
> 
> If thats the case, And let say we have an M3UA Endpoint at SGP node as E1 [
> IP-A, 2905 ] & this SGP has associations with 100 ASP's .   And in this case
> if i do ADDIP(IPb) on this Endpoint E1, than I have a major problem because
> i have to send 100 ASCONF chunks one-by-one on all associations, which seems
> very difficult & cumbersome taking in to consideration the race conditions
> and scenarios mentioned in my previous mails.
> 
> Any comments ???
> 
> thanks
> sandeep
> 
> > -----Original Message-----
> > From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> > Sent: Monday, April 29, 2002 1:58 PM
> > To: Qiaobing Xie
> > Cc: Qiaobing Xie; Tuexen Michael; Ivan.Arias-Rodriguez@nokia.com;
> > s_balani@trillium.com; mo_khan@trillium.com; tsvwg@ietf.org
> > Subject: Re: [Tsvwg] End point and association Address conflict
> >
> >
> > Qiaobing Xie wrote:
> > >
> > > ...
> > > > I don't agree with this rule.. just because a peer in a UDP model
> > > > endpoint on one of its assocaitions is NOT capable of add-ip why
> > > > should the rest be stuck. Yes you have an abnormality here in
> > > > that that peer COULD send a INIT to the new address..
> > again I repeat:
> > > > 1) This is such a corner case it is almost silly.
> > > > 2) If it happens it is not that hard to handle... Looking at
> > > >    the code Peter and I wrote you would end up finding
> > the endpoint
> > > >    (since the address WAS added there) and you would then LOOK at
> > > >    its associations. This would then appear to be a
> > RESTART where the
> > > >    peer added addresses. You would then SEND IT an ABORT
> > saying restart
> > > >    while adding addresses. This would leave the other association
> > > >    un-touched and would elimiate the error on the peers part...
> > >
> > > Randy,
> > >
> > > Let's walk through this example in detail:
> > > Say, local endpoint E1 has two associations to different peers,
> > >
> > >      E1             A1     E2
> > > [IPA,IPB:Port1] <-------> [...]
> > >      E1             A2     E3
> > > [IPA,IPB:Port1] <-------> [IPZ:Port2]
> > >
> > > We do add-ip (IPC) to E1. After sending ASCONF to both E2 and E3, we
> > > succeeded with E2 but failed on E3.
> > >
> > >      E1'               A1'     E2
> > > [IPA,IPB,IPC:Port1] <-------> [...]
> > >      E1'               A2'     E3
> > > [IPA,IPB,IPC:Port1] <-------> [IPZ:Port2]  /* E3 rejected
> > add-ip and */
> > >                                       /* hence knows nothg
> > about IPC */
> > >
> > > My understanding is that you would like to allow this.
> > >
> > > Now, several things could happen,
> > >
> > > 1) E3 may get a packet from E1' with src addr=IPC. This packet will
> > > becomes an OOTB to E3 and E3 will send back an ABORT to kill the
> > > ill-fated A2'. This is ok.
> > >
> >
> > Why would E1' use IPC... in this case it knows that it is NOT
> > part of the association. So it MUST NEVER source things from
> > IPC as long as this holds true .. no problem.. only if
> > E1's implementation is so bad it sources from IPC that a
> > problem will occur.
> >
> >
> > > 2) E3 may send an INIT to IPC/Port1. Is this the corner case you
> > > described above? If so, I can see E1'1 will find the
> > association A2' and
> > > take this E3 INIT as a restart, but I don't see why E1'
> > will consider it
> > > as a "restart with added address" (E3 didn't add any
> > address, it still
> > > uses only IPZ).
> >
> > Yes.. this is a very corner case..
> >
> >  A) E3 must be very confused to want two associations.. so in
> >     my mind this is a corner case.
> >
> >  B) We would need to expand the rules that you must also check
> >     the your addresses in a COOKIE-ECHO to see if you have
> >     hit such a collision.. if so send an ABORT...
> >
> > I don't see this is a problem.. ABORTING your funny association
> > with E3 should leave A2 alone and all should be well.
> >
> > It seems to me we need to add this new rule for ADD-IP.
> >
> > R
> >
> > --
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> >


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 20:48:08 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA24469
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 20:48:08 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA06755
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 20:48:11 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA05648;
	Mon, 29 Apr 2002 20:22:46 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA05621
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 20:22:43 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA24192
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 20:22:40 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3U0MKH60740;
	Mon, 29 Apr 2002 19:22:21 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCDE3BC.A623629@stewart.chicago.il.us>
Date: Mon, 29 Apr 2002 19:22:20 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Qiaobing Xie <qxie1@email.mot.com>
CC: Michael Tuexen <Michael.Tuexen@icn.siemens.de>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, s_balani@trillium.com,
        mo_khan@trillium.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <875228B9-5BAC-11D6-ACAA-0030654C1AB6@icn.siemens.de> <3CCDDF38.6EA2990E@email.mot.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Qiaobing Xie wrote:
> 
> Hi, Michael,
> 
> > on local host         on remote host(s)
> > A1: E1 <-----------------> E3
> > A2: E1 <---------------------> E4
> > A3: E1 <-------------------------> E5
> ...
> > > Rule 2: when adding a transport to a local endpoint that has
> > > associations to multiple peers (ie, a fan-out config), one MUST send an
> > > add-ip ASCONF to each of those peers. And the new transport can only be
> > > used until ALL the peers ack successfully. If anyone of the peers denies
> > > the add-ip then you can't add it. So, if you add a new transport to E1
> > > above, you have to send ASCONF to both E3 and E4 and have to get a
> > > positive ack from both of them.
> > >
> > I do NOT like this rule. How do you handle the case if E3 and E4 support
> > AddIP, but E5 does not. Do you say that you can not add an address to E1
> > because one of the peers does not support AddIP? for deployment this means
> > that AddIP only be used if all implementations support this...
> 
> An alternative is to find a way to abort E5 and force it to restart with
> E1. Maybe we should introduce an ABORT(error code=please restart with
> me).
> 
> > Why not add it to E1, but it is not allowed to be used as a source
> > address for A5.
> 
> This will split E1 into two logical endpoints: E1-with-new-addr, and
> E1-without-new-addr. And the two logically different endpoints will end
> up sharing all original addresses. This breaks the basic RFC2309
> definition/rule that says:
>     "A transport address used by an SCTP endpoint must not be used
>      by another SCTP endpoint.  In other words, a transport address
>      is unique to an SCTP endpoint."


I think we are still out of wack on definition of an endpoint
here. I don't see how the associations an endpoint E1 has can
split E1 into two logical endpoints.. it is STILL E1 and it
just has an association amongst its three ones that can't use
one of its addresses. This is NO different than E1 having
bound ALL IPv6 addresses Link local, Site Local and Global and
then setting up an association with a peer that is NOT on its
Site and NOT on its link.. and at the same time setting up an
association with a peer on the local link that is ALSO in the
same site. So you end up with:

<Endpoint Port:2222 IPv6G, IPv6S, IPV6LL>

E1(Port:2222 IPv6G)<------A1------------>PeerEX(Port:3333 IPv6G)
and
E1(Port:2222 IPv6LL, IPv6S, IPv6G)<------A2------->PeerIN(Port:4444
                                                   IPv6G,IPv6S,IPvLL)

Is E1 split by this... NO!!! 

You just have the case that E1 has different addresses active on one
association versus another... this is no different than

<Endpoint IP-X:Port 2222>
E1(Port:2222 IP-X)<-------------A10---------------->Peer1(Port:3333
IP-A)
E1(Port:2222 IP-X)<-------------A11---------------->Peer2(Port:4444
IP-B)


Now if I do a Add of IP-Y to the endpoint and
then find that peer1 does asconf and peer2 does not I end up with:

we have

<Endpoint IP-X,IP-Y: Port 2222>
E1(Port:2222 IP-X,IP-Y)<-----------A10--------------->Peer1(Port:3333
IP-A)
E1(Port:2222 IP-X)<----------------A11--------------->Peer2(Port:4444
IP-B)

A11 just is using a subset of the addresses in the same way the 
IPv6 A1 did the same thing... 

This did NOT magically make E1 into E1', Endpoint 1 is still Endpoint 1
It owns the addresses it has bound .. period...

When you attempt to bind IP-Y to the endpoint for port 2222 you must
verify that NO OTHER endpoint has it bound.. as long as that is true..
No problem... You add that one to the endpoint... and if some
association can only use a subset of the addresses so be it...

Now as to the corner case.. I think I have shown that this is NOT
a problem... proper coding will take care of this issue..

Do we need some documentation somewhere? Maybe...

R
-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 20:48:10 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA24487
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 20:48:10 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA06770
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 20:48:12 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA05336;
	Mon, 29 Apr 2002 20:17:02 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA04064
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 20:00:29 -0400 (EDT)
Received: from motgate2.mot.com (motgate2.mot.com [136.182.1.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA23828
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 20:00:19 -0400 (EDT)
Received: [from pobox2.mot.com (pobox2.mot.com [136.182.15.8]) by motgate2.mot.com (motgate2 2.1) with ESMTP id RAA22193; Mon, 29 Apr 2002 17:00:14 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox2.mot.com (MOT-pobox2 2.0) with ESMTP id RAA01875; Mon, 29 Apr 2002 17:00:09 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3U006b00447;
	Mon, 29 Apr 2002 19:00:07 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id TAA22002;
	Mon, 29 Apr 2002 19:00:17 -0500 (CDT)
Message-ID: <3CCDDF38.6EA2990E@email.mot.com>
Date: Mon, 29 Apr 2002 19:03:04 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Michael Tuexen <Michael.Tuexen@icn.siemens.de>
CC: Qiaobing Xie <Qiaobing.Xie@motorola.com>, Ivan.Arias-Rodriguez@nokia.com,
        randall@stewart.chicago.il.us, s_balani@trillium.com,
        mo_khan@trillium.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <875228B9-5BAC-11D6-ACAA-0030654C1AB6@icn.siemens.de>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi, Michael,

> on local host         on remote host(s)
> A1: E1 <-----------------> E3
> A2: E1 <---------------------> E4
> A3: E1 <-------------------------> E5
...
> > Rule 2: when adding a transport to a local endpoint that has
> > associations to multiple peers (ie, a fan-out config), one MUST send an
> > add-ip ASCONF to each of those peers. And the new transport can only be
> > used until ALL the peers ack successfully. If anyone of the peers denies
> > the add-ip then you can't add it. So, if you add a new transport to E1
> > above, you have to send ASCONF to both E3 and E4 and have to get a
> > positive ack from both of them.
> >
> I do NOT like this rule. How do you handle the case if E3 and E4 support
> AddIP, but E5 does not. Do you say that you can not add an address to E1 
> because one of the peers does not support AddIP? for deployment this means 
> that AddIP only be used if all implementations support this...

An alternative is to find a way to abort E5 and force it to restart with
E1. Maybe we should introduce an ABORT(error code=please restart with
me).
 
> Why not add it to E1, but it is not allowed to be used as a source
> address for A5.

This will split E1 into two logical endpoints: E1-with-new-addr, and
E1-without-new-addr. And the two logically different endpoints will end
up sharing all original addresses. This breaks the basic RFC2309
definition/rule that says: 
    "A transport address used by an SCTP endpoint must not be used
     by another SCTP endpoint.  In other words, a transport address 
     is unique to an SCTP endpoint."

-Qiaobing


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 20:51:06 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA24525
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 20:51:06 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA06860
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 20:50:56 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA05378;
	Mon, 29 Apr 2002 20:17:20 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA04535
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 20:03:15 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA23878
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 20:03:12 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3U02WH60695;
	Mon, 29 Apr 2002 19:02:33 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCDDF18.9DDD2030@stewart.chicago.il.us>
Date: Mon, 29 Apr 2002 19:02:32 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Qiaobing Xie <qxie1@email.mot.com>
CC: Jon Grimm <jgrimm@austin.ibm.com>, Ivan.Arias-Rodriguez@nokia.com,
        bidulock@openss7.org, Michael.Tuexen@icn.siemens.de,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>, s_balani@trillium.com,
        mo_khan@trillium.com, h_pillai@trillium.com, jheitz@lucent.com,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFB4@esebe017.NOE.Nokia.com> <3CCD844A.DB1BC960@austin.ibm.com> <3CCD9267.6456A90E@stewart.chicago.il.us> <3CCDA2AD.CB11491@austin.ibm.com> <3CCDB210.E8FA4663@stewart.chicago.il.us> <3CCDD167.8A0DBC43@email.mot.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

A comment in line and then some more thoughts..

Qiaobing Xie wrote:
> 
> Hi, Randy,
> 
> I think we are still using slightly different endpoint definitions (as
> you also realized below).
> 
> ...
> >
> > I still view it as adding to an endpoint but I think we must
> > re-think about this... I put this down in a private email discussion
> > but I think I will put it out here now ... even though the
> > thoughts are only partially baked :>
> >
> > We have basically 4 types of endpoints/sockets (note I think of
> > an endpoint as a socket aka open FD).
> 
> I think of an endpoint as an abstract logical entity (as dryly defined
> in RFC2960). I see three levels of address sets involved here:
> 
> Level 0: host addr set "S" = all the network addresses assigned to a
> host (host is logically represented by a single OS instance). A host can
> harbor multiple sockets/open FDs.
> 
> Level 1: addr set bound to a socket/open FD "s" = a subset of S bound to
> the FD (this is the logical representation of an SCTP stack). An open FD
> can harbor multiple endpoints. But some implementation, for simplicity
> reason, may support only one endpoint.
> 
> Level 2: addr set of an endpoint "e" = yet a subset of "s". This is the
> set of addresses actually used (or to be used) in an association.
> 

I don't get your Level 2 definition.. it is unclear to me
what you are saying...

> In summary, e <= s <= S.
> 
> All my previous discussions were based on the "e" definition (including
> the rules I proposed).
> Therefore, the following statement from Ivan
> 
>   - An association, can use a subset of the source addresses of the
>     endpoint to which it belongs.
> 
> becomes correct if Ivan's "endpoint" here implies a socket/open FD, but
> incorrect if the "endpoint" here is defined as "e" above.
> 
> I will think more about the socket implications.

This may help.. think on what I have said about socket descriptors (or
fds if you like) and what type of socket api model we have..

Now I have done a bit more thinking on this. The corner case that
we were worried about was this (if I remember right):


Host A                                          Host Z
IP-A/IP-B                                       IP-Z
(An UDP model Endpoint
 Listening on Port 2222)
        IP-A:Port-2222<-----------------INIT(From IP-Z:Port 3333)
        ----------INIT-ACK(IP-A,IP-B:Port 2222)-------->
        <--------------COOKIE-ECHO----------------------
        ----------------COOKIE-ACK---------------------->

            Now at this point A-1 is up.

Now along the line we do

ifconfg New-Intf IP-C

and magically do

--------------ASCONF(ADD-IP:IP-C)------------------------>
    <---------------------OperationalError(Unknown Chunk ASCONF)-----

So here we have a potential problem...

                    now our silly peer does

 <------------------------IP-C:2222 INIT(IP-Z:Port 3333)

What do we do?

 --------INIT-ACK(IP-A,IP-B,IP-C: Port 2222)---------->IP-Z:Port
3333--->

                 
Now what we have forgotten about is what our silly endpoint is going
to do. 

The SCTP stack on Host-Z must look to see if a collision is
happening.. this is required in RFC2960. So what happens will depend
on where the INIT-ACK is sourced from..

If by chance its source is IP-A or IP-B ... guess what? the lookup
at Host-Z will return the OLD association which is NOT in a COOKIE-WAIT
state.. so what will Host-Z doe? Discard the INIT-ACK... This will
happen again (i.e. the INIT-timer will expire on Host-Z and this will
then repeat until we fail).... No problem..

Now what happens if by chance it IS sourced from IP-C? We find the
newly forming association. Now it goes throughs its procedures 
to attempt to gather all the valuable goodies out of the INIT-ACK.
In the process it must load its peers addresses to the association.
And here is where it SHOULD detect the other association. I know
in our implementation one of the first things I do when loading
an address is do a lookup on the address I am loading.. I want
to make sure it is in the association only ONCE. In the process
I get back the association ... one of the checks I do is verify
that it is the same association that I am working on loading 
address to .. if not.. guess what ? I detect the error and
bail out.. dropping the INIT-ACK... so .. same result.. we continue
to send out INIT's and drop the result.. thus the IP-C association
just never forms..

Now is this clear in any document that you must look for addresses
as you load them.. no its not specified since it is an implementation
detail. I will admit it is an important one.. Should it be documented
that you SHOULD attempt to detect this? Maybe but I am not sure 
where this would fit..

But in the end my basic premise is that if the implementation is
coded right the silly endpoint will self-correct and NOT bring up
this strange association.

I don't think we can dis-allow adding an new address... otherwise
a true restarting association could not come up if I just added
an address until it aborted ... so I don't like that idea...

Let me also note .. say my implementation hits a peer that does
NOT catch this.. guess what.. the same code is going to run to
load the addresses from the cookie.. It too will find this 
double association setup and drop the cookie...

So my contention is that this corner case is NOT an issue if
the address loading is done properly by the implementation...



R


> 
> >
> > Now we have:
> >
> > 1) A UDP model socket/listener that A) accepts new associations and
> >    has some number of associations under it (fanned out is the term
> >    I think Qiaobing used).
> >
> > 2) A TCP model listener. This guy will have NO associations under
> >    it accept maybe onces waiting to be accepted ... nothing can
> >    happen on those though.. until AFTER they are accepted.
> >
> > 3) Accepted and or peel'ed off ones.. these are a single association.
> >
> > 4) A TCP model connect() caller. This guy can NOT accept any.. but is
> >    going to have a single "association" to a peer... note this may
> >    be (in the end) to a <1> or <3> type guy.. it does not matter.
> >
> > Now lets look at each case..
> >
> > First, an easy one, <3> is something that can NEVER get any new
> > associations. INIT's don't go to it... they go to whoever was the
> > listner... so in this case this association COULD have add-ip go
> > against it. .thoughts/notes...{I believe in our implementation of add-ip
> > this
> >                                guy MAY auto-add IP if it is setup via an
> >                                sockopt to do it.. default may be on
> > especially
> >                                if the parent (listener) is bound all
> >                                }
> >
> > Now lets look at another easy one.. <4> the connector.. this guy
> > IS NOT LISTENING.. it TOO can't get a INIT in.. an ABORT() would be
> > sent back... At the termination of its association, however, it MAY
> > do a listen() call..  so here we would ADD the address to the endpoint
> > (if it was turned on socketopt wise to do so..) and the attempt if
> > possible to add it to the peer.
> >
> > Now this leaves case <1> and <2>. The only real difference between
> > these guys is that we have one that has a cluster of associations
> > tied to it.. the other does NOT. Now we here ADD the address to the
> > endpoint. For <2> we are done.. when the next INIT comes in the new
> > association will form... and we are ok... For <1> new associations
> > will use the new address.. (all of this assumes that the option
> > to add automatically is enabled.. for bind-all this is probably
> > the default.. it may not be for sub-set bound). Now for the
> > existing "fan" of associations we would individually attempt to
> > add the address to them..
> >
> > Now for the big question .. what happens if we get our silly endpoint
> > (peer) that sends an INIT while holding an association to us (either
> > a TCP one or a fanned out one). Well some observations:
> >
> > a) First you must be able to find the old associations, if you
> >    can't then your INIT collision and RESTART code is NOT working.
> >
> > b) When you find the "old" associations it is going to look like
> >    a restart condition... especially for <1> since you locate the
> >    endpoint (and you have added the address here hopefully) and then
> >    you look for the association via one of the peers addresses. For the
> >    TCP model you also must switch the socket you are associated with
> >    in this hunt... hmmm lets look at both sides..
> >     - If you find it then you would actually go through the
> >       restart scenario as long as the peer used the same addresses
> >       it would appear to be a restart.. this would be very interesting
> >       and of course the peer would get very messed up I bet... but its
> >       a silly endpoint anyway.. in the end the old association would
> >       die, one way or another... Now if we add a restriction that we
> >       can't add any new addresses on either side (aka like our current
> >       peer can't add new addresses) then no problem.. we would ABORT()
> >       with new addresses...
> >     - If you can't find it then what would happen is when the COOKIE
> >       arrives you could since you could look at all your addresses
> >       and validate that a collision is not at hand... and then send
> >       an abort...
> >
> >    Pesonnely I prefer the no new addresses on either side during a
> >    collision.. this solves the problem as long as you can locate your
> >    associations.. which I think you should be able to..
> >
> >  Well .. there are some of my 1/2 baked thoughts... need to mull on
> >  them a bit more..
> >
> > R
> >
> > --
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 20:57:33 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA24632
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 20:57:33 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA07114
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 20:57:36 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA05150;
	Mon, 29 Apr 2002 20:16:18 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA20707
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 16:51:10 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA19416
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 16:51:05 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3TKoPH60362;
	Mon, 29 Apr 2002 15:50:25 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCDB210.E8FA4663@stewart.chicago.il.us>
Date: Mon, 29 Apr 2002 15:50:25 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Jon Grimm <jgrimm@austin.ibm.com>
CC: Ivan.Arias-Rodriguez@nokia.com, bidulock@openss7.org,
        Michael.Tuexen@icn.siemens.de, qxie1@email.mot.com,
        s_balani@trillium.com, Qiaobing.Xie@motorola.com, mo_khan@trillium.com,
        h_pillai@trillium.com, jheitz@lucent.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFB4@esebe017.NOE.Nokia.com> <3CCD844A.DB1BC960@austin.ibm.com> <3CCD9267.6456A90E@stewart.chicago.il.us> <3CCDA2AD.CB11491@austin.ibm.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Jon Grimm wrote:
> 
> Thanks for the remarks, Randy.
>  I have a few comments still below.  Thanks, Jon
> 
> Randall Stewart wrote:
> >
> > Jon:
> >
> > a couple of comments :>
> >
> > Jon Grimm wrote:
> > >
> > > Ivan/All
> > >
> > >         I think something that needs to come out of this discussion are
> > > scenarios and mandatory behavior in those scenarios (or why flexibility
> > > is allowed if not mandatory behavior).  Its a little disturbing that a
> > > similar 'what is an endpoint?' discussion seems to pop up again every so
> > > often.  (e.g.bindx discussion last Sept in
> > > ftp://ftp.ietf.org/ietf-mail-archive/tsvwg/2001-09.mail ) and still does
> > > not seem to be completely clear even to those following SCTP closely
> > > over the last year.
> >
> > I agree.. it is a discussion that seems to NEVER end :>
> >
> > >
> > > For each of the following scenarios,
> > >
> > > 1) A new address is to be added an endpoint. (manual)
> > > 2) An address is to be deleted from an endpoint.  (manual)
> > > 3) A new NIC/address is configured and an endpoint is bound to all
> > > (auto)
> > > 4) An address bound is brought down (no longer available) (auto).
> > > 5) An address bound is brought down for maintenance and then brought
> > > on-line again. (auto)
> > >
> > > consider each of the following subcases (assuming multiple associations
> > > under an ep).
> > >
> > > a) All Peers have ADD-IP.
> > > b) No Peers have ADD-IP.
> > > c) A subset of Peers support ADD-IP.
> > >
> > > What happens?
> > >
> >
> > Where are you proposing putting these examples? I think some
> > text is needed for clearification but I am not sure that a protocol
> > spec is where you illustrate all these different examples...
> 
> In some I-D,of course.   I see you are considering the Sockets API I-D,
> yes this could work for me as an appendix there.  Given this latest
> round of discussion, folks may wildly disagree, but I'm coming from the
> angle that I just want it captured.  ;-)
> 
> >
> > > Another comment:
> > > Dropping down a little bit into the Sockets API for illustration:
> > >
> > > Depending on one's interpretation of endpoint, what is the endpoint for
> > > the association of a peeled-off UDP style socket or for that matter an
> > > 'accepted' TCP-style socket.  If the strict interpretation of the
> > > endpoint is kept.. wouldn't an ADD-IP on the child sockets (peeled off
> > > or accept TCP) affect the local endpoint of the parent socket?  Or vise
> > > versa, should an ADD-IP on a TCP-style listening socket, affect the
> > > accepted sockets?  If yes, was this expected?  If no, does the original
> >
> > In my view the ADD-IP is done against the listening type endpoint. So
> > in the case of connected and/or peeled off associations they are
> > a different critter. In some ways they have formed a "new" endpoint. But
> > a interesting thing about them is they CAN NOT listen. So maybe in
> > this case Brian's adding an address to the association is the right
> > approach. They can NEVER receive a new association.. only the single
> > listner (in the TCP or UDP case) can do that.
> >
> > So a peeled off or "connected" socket with only one association may
> > well be a candidate to do an add-ip on an association basis.. but not
> > on the listening endpoint or a UDP model grouping of associations...
> >
> 
> Has this gone full circle then?   This sounds more like you are always
> doing add-ip on associations.   Or you are really always doing add-ip on
> (Brian's or Ivan's ??) individual communication endpoints that belong to
> a grouping (group of 1 (peeled off) or many (udp style))?
> 

I still view it as adding to an endpoint but I think we must
re-think about this... I put this down in a private email discussion
but I think I will put it out here now ... even though the
thoughts are only partially baked :>


We have basically 4 types of endpoints/sockets (note I think of
an endpoint as a socket aka open FD).

Now we have:

1) A UDP model socket/listener that A) accepts new associations and
   has some number of associations under it (fanned out is the term
   I think Qiaobing used).

2) A TCP model listener. This guy will have NO associations under
   it accept maybe onces waiting to be accepted ... nothing can
   happen on those though.. until AFTER they are accepted.

3) Accepted and or peel'ed off ones.. these are a single association.

4) A TCP model connect() caller. This guy can NOT accept any.. but is
   going to have a single "association" to a peer... note this may
   be (in the end) to a <1> or <3> type guy.. it does not matter.

Now lets look at each case..

First, an easy one, <3> is something that can NEVER get any new
associations. INIT's don't go to it... they go to whoever was the
listner... so in this case this association COULD have add-ip go
against it. .thoughts/notes...{I believe in our implementation of add-ip
this
                               guy MAY auto-add IP if it is setup via an
                               sockopt to do it.. default may be on
especially
                               if the parent (listener) is bound all 
                               }

Now lets look at another easy one.. <4> the connector.. this guy
IS NOT LISTENING.. it TOO can't get a INIT in.. an ABORT() would be
sent back... At the termination of its association, however, it MAY
do a listen() call..  so here we would ADD the address to the endpoint
(if it was turned on socketopt wise to do so..) and the attempt if
possible to add it to the peer.

Now this leaves case <1> and <2>. The only real difference between
these guys is that we have one that has a cluster of associations
tied to it.. the other does NOT. Now we here ADD the address to the
endpoint. For <2> we are done.. when the next INIT comes in the new
association will form... and we are ok... For <1> new associations
will use the new address.. (all of this assumes that the option
to add automatically is enabled.. for bind-all this is probably
the default.. it may not be for sub-set bound). Now for the
existing "fan" of associations we would individually attempt to
add the address to them..

Now for the big question .. what happens if we get our silly endpoint
(peer) that sends an INIT while holding an association to us (either
a TCP one or a fanned out one). Well some observations:

a) First you must be able to find the old associations, if you
   can't then your INIT collision and RESTART code is NOT working.

b) When you find the "old" associations it is going to look like
   a restart condition... especially for <1> since you locate the
   endpoint (and you have added the address here hopefully) and then
   you look for the association via one of the peers addresses. For the
   TCP model you also must switch the socket you are associated with
   in this hunt... hmmm lets look at both sides..
    - If you find it then you would actually go through the
      restart scenario as long as the peer used the same addresses
      it would appear to be a restart.. this would be very interesting
      and of course the peer would get very messed up I bet... but its
      a silly endpoint anyway.. in the end the old association would 
      die, one way or another... Now if we add a restriction that we
      can't add any new addresses on either side (aka like our current
      peer can't add new addresses) then no problem.. we would ABORT()
      with new addresses...
    - If you can't find it then what would happen is when the COOKIE
      arrives you could since you could look at all your addresses
      and validate that a collision is not at hand... and then send
      an abort...

   Pesonnely I prefer the no new addresses on either side during a
   collision.. this solves the problem as long as you can locate your
   associations.. which I think you should be able to..

 Well .. there are some of my 1/2 baked thoughts... need to mull on
 them a bit more..

R               

--
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 21:23:37 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA25002
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 21:23:36 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id VAA08476
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 21:23:39 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA06901;
	Mon, 29 Apr 2002 20:51:11 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA06873
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 20:51:09 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-230.dallas.net [209.44.42.230])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA24524
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 20:51:03 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id TAA32318;
	Mon, 29 Apr 2002 19:50:39 -0500
Date: Mon, 29 Apr 2002 19:50:39 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Randall Stewart <randall@stewart.chicago.il.us>
Cc: Qiaobing Xie <qxie1@email.mot.com>,
        Michael Tuexen <Michael.Tuexen@icn.siemens.de>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, s_balani@trillium.com,
        mo_khan@trillium.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
Message-ID: <20020429195039.J15434@openss7.org>
Reply-To: bidulock@openss7.org
References: <875228B9-5BAC-11D6-ACAA-0030654C1AB6@icn.siemens.de> <3CCDDF38.6EA2990E@email.mot.com> <3CCDE3BC.A623629@stewart.chicago.il.us>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <3CCDE3BC.A623629@stewart.chicago.il.us>; from randall@stewart.chicago.il.us on Mon, Apr 29, 2002 at 07:22:20PM -0500
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Randall,

On Mon, 29 Apr 2002, Randall Stewart wrote:

[snip]
> 
> When you attempt to bind IP-Y to the endpoint for port 2222 you must
> verify that NO OTHER endpoint has it bound.. as long as that is true..

There is no such restriction in TCP.  It is quite permissible (and
falls under the category of local port reuse) to have the following
binding:

    E1 [IPA/PortX] ---
    E2 [IPA/PortX] ---

In fact these may start off life as:

    E1 [IPA/Port0] --
    E2 [IPA/Port0] --

And when they connect, they autobind with local port reuse and are
permitted to become:

    E1 [IPA/PortX] ------------------  [IPB/PortY]
    E2 [IPA/PortX] ------------------  [IPB/PortZ]

(These ports were never listening.)

Why does SCTP require that this not happen when it is not only
permissible but a common port reuse strategy in TCP?

--brian

> No problem... You add that one to the endpoint... and if some
> association can only use a subset of the addresses so be it...
> 
> Now as to the corner case.. I think I have shown that this is NOT
> a problem... proper coding will take care of this issue..
> 
> Do we need some documentation somewhere? Maybe...
> 
> R
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 22:09:34 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA26492
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 22:09:34 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id WAA11165
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 22:09:37 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id VAA09504;
	Mon, 29 Apr 2002 21:42:51 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA07140
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 20:59:28 -0400 (EDT)
Received: from ftpbox.mot.com (ftpbox.mot.com [129.188.136.101])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA24669
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 20:59:25 -0400 (EDT)
Received: [from pobox2.mot.com (pobox2.mot.com [136.182.15.8]) by ftpbox.mot.com (ftpbox 2.1) with ESMTP id RAA13392; Mon, 29 Apr 2002 17:59:27 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox2.mot.com (MOT-pobox2 2.0) with ESMTP id RAA12377; Mon, 29 Apr 2002 17:59:27 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3U0xNb03699;
	Mon, 29 Apr 2002 19:59:23 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id TAA22051;
	Mon, 29 Apr 2002 19:59:33 -0500 (CDT)
Message-ID: <3CCDED1D.EFC1E6F0@email.mot.com>
Date: Mon, 29 Apr 2002 20:02:21 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Randall Stewart <randall@stewart.chicago.il.us>
CC: Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Jon Grimm <jgrimm@austin.ibm.com>, Ivan.Arias-Rodriguez@nokia.com,
        bidulock@openss7.org, Michael.Tuexen@icn.siemens.de,
        s_balani@trillium.com, mo_khan@trillium.com, h_pillai@trillium.com,
        jheitz@lucent.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFB4@esebe017.NOE.Nokia.com> <3CCD844A.DB1BC960@austin.ibm.com> <3CCD9267.6456A90E@stewart.chicago.il.us> <3CCDA2AD.CB11491@austin.ibm.com> <3CCDB210.E8FA4663@stewart.chicago.il.us> <3CCDD167.8A0DBC43@email.mot.com> <3CCDDF18.9DDD2030@stewart.chicago.il.us>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

R, 

A couple of comments...

Randall Stewart wrote:
...
> Now I have done a bit more thinking on this. The corner case that
> we were worried about was this (if I remember right):
> 
> Host A                                          Host Z
> IP-A/IP-B                                       IP-Z
> (An UDP model Endpoint
>  Listening on Port 2222)
>         IP-A:Port-2222<-----------------INIT(From IP-Z:Port 3333)
>         ----------INIT-ACK(IP-A,IP-B:Port 2222)-------->
>         <--------------COOKIE-ECHO----------------------
>         ----------------COOKIE-ACK---------------------->
> 
>             Now at this point A-1 is up.
> 
> Now along the line we do
> 
> ifconfg New-Intf IP-C
> 
> and magically do
> 
> --------------ASCONF(ADD-IP:IP-C)------------------------>
>     <---------------------OperationalError(Unknown Chunk ASCONF)-----
> 
> So here we have a potential problem...

At this point, would our local stack (I am avoiding to use endpoint
here) actually start to use IP-C as the source addr in any outbound
packets in the current association? If so, that would trigger an abort
from the peer and kill the current association. If not, then strictly
speaking, IP-C has not yet been added to the local **endpoint** that the
current associaiton is using. It has only been made available to the
local stack for future use. 

> 
>                     now our silly peer does
> 
>  <------------------------IP-C:2222 INIT(IP-Z:Port 3333)
> 
> What do we do?
> 
>  --------INIT-ACK(IP-A,IP-B,IP-C: Port 2222)---------->IP-Z:Port
> 3333--->
> 

When the local stack does this, it starts to manage two logically
different local endpoints - the one used by the old association that has
only IP-A and IP-B; the other is intended for the new association, as
advertised in this INIT-ACK, that has IP-A, IP-B, and IP-C.

To have multiple logical endpoints here is not a problem. But the
problem is that they are sharing IP-A and IP-B. This not only goes
against RFC2960 definition, but also becomes the root cause that opens
the door for the "corner case" and potential danger of other "duplicate
association" cases.

Of cause, as you showed below, if a solid implementation is vigorously
checking and double-checking for dup associations at multiple points
during the initiation process, you may eventually weed them out. But I
am not sure about just relying on the implementation to catch dup
associations.

-Qiaobing

> 
> Now what we have forgotten about is what our silly endpoint is going
> to do.
> 
> The SCTP stack on Host-Z must look to see if a collision is
> happening.. this is required in RFC2960. So what happens will depend
> on where the INIT-ACK is sourced from..
> 
> If by chance its source is IP-A or IP-B ... guess what? the lookup
> at Host-Z will return the OLD association which is NOT in a COOKIE-WAIT
> state.. so what will Host-Z doe? Discard the INIT-ACK... This will
> happen again (i.e. the INIT-timer will expire on Host-Z and this will
> then repeat until we fail).... No problem..
> 
> Now what happens if by chance it IS sourced from IP-C? We find the
> newly forming association. Now it goes throughs its procedures
> to attempt to gather all the valuable goodies out of the INIT-ACK.
> In the process it must load its peers addresses to the association.
> And here is where it SHOULD detect the other association. I know
> in our implementation one of the first things I do when loading
> an address is do a lookup on the address I am loading.. I want
> to make sure it is in the association only ONCE. In the process
> I get back the association ... one of the checks I do is verify
> that it is the same association that I am working on loading
> address to .. if not.. guess what ? I detect the error and
> bail out.. dropping the INIT-ACK... so .. same result.. we continue
> to send out INIT's and drop the result.. thus the IP-C association
> just never forms..
> 
> Now is this clear in any document that you must look for addresses
> as you load them.. no its not specified since it is an implementation
> detail. I will admit it is an important one.. Should it be documented
> that you SHOULD attempt to detect this? Maybe but I am not sure
> where this would fit..
> 
> But in the end my basic premise is that if the implementation is
> coded right the silly endpoint will self-correct and NOT bring up
> this strange association.
> 
> I don't think we can dis-allow adding an new address... otherwise
> a true restarting association could not come up if I just added
> an address until it aborted ... so I don't like that idea...
> 
> Let me also note .. say my implementation hits a peer that does
> NOT catch this.. guess what.. the same code is going to run to
> load the addresses from the cookie.. It too will find this
> double association setup and drop the cookie...
> 
> So my contention is that this corner case is NOT an issue if
> the address loading is done properly by the implementation...
> 
> R
>


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 22:21:17 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA26608
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 22:21:17 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id WAA11389
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 22:21:20 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id VAA10047;
	Mon, 29 Apr 2002 21:53:55 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id VAA09595
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 21:45:27 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA26122
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 21:45:22 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3U1fLH60861;
	Mon, 29 Apr 2002 20:41:22 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCDF641.56A4C9AD@stewart.chicago.il.us>
Date: Mon, 29 Apr 2002 20:41:21 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Qiaobing Xie <qxie1@email.mot.com>
CC: Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Jon Grimm <jgrimm@austin.ibm.com>, Ivan.Arias-Rodriguez@nokia.com,
        bidulock@openss7.org, Michael.Tuexen@icn.siemens.de,
        s_balani@trillium.com, mo_khan@trillium.com, h_pillai@trillium.com,
        jheitz@lucent.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFB4@esebe017.NOE.Nokia.com> <3CCD844A.DB1BC960@austin.ibm.com> <3CCD9267.6456A90E@stewart.chicago.il.us> <3CCDA2AD.CB11491@austin.ibm.com> <3CCDB210.E8FA4663@stewart.chicago.il.us> <3CCDD167.8A0DBC43@email.mot.com> <3CCDDF18.9DDD2030@stewart.chicago.il.us> <3CCDED1D.EFC1E6F0@email.mot.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Qiaobing Xie wrote:
> 
> R,
> 
> A couple of comments...
> 
> Randall Stewart wrote:
> ...
> > Now I have done a bit more thinking on this. The corner case that
> > we were worried about was this (if I remember right):
> >
> > Host A                                          Host Z
> > IP-A/IP-B                                       IP-Z
> > (An UDP model Endpoint
> >  Listening on Port 2222)
> >         IP-A:Port-2222<-----------------INIT(From IP-Z:Port 3333)
> >         ----------INIT-ACK(IP-A,IP-B:Port 2222)-------->
> >         <--------------COOKIE-ECHO----------------------
> >         ----------------COOKIE-ACK---------------------->
> >
> >             Now at this point A-1 is up.
> >
> > Now along the line we do
> >
> > ifconfg New-Intf IP-C
> >
> > and magically do
> >
> > --------------ASCONF(ADD-IP:IP-C)------------------------>
> >     <---------------------OperationalError(Unknown Chunk ASCONF)-----
> >
> > So here we have a potential problem...
> 
> At this point, would our local stack (I am avoiding to use endpoint
> here) actually start to use IP-C as the source addr in any outbound
> packets in the current association? If so, that would trigger an abort
> from the peer and kill the current association. If not, then strictly
> speaking, IP-C has not yet been added to the local **endpoint** that the
> current associaiton is using. It has only been made available to the
> local stack for future use.
> 
> >
> >                     now our silly peer does
> >
> >  <------------------------IP-C:2222 INIT(IP-Z:Port 3333)
> >
> > What do we do?
> >
> >  --------INIT-ACK(IP-A,IP-B,IP-C: Port 2222)---------->IP-Z:Port
> > 3333--->
> >
> 
> When the local stack does this, it starts to manage two logically
> different local endpoints - the one used by the old association that has
> only IP-A and IP-B; the other is intended for the new association, as
> advertised in this INIT-ACK, that has IP-A, IP-B, and IP-C.
> 

No it is managing two different assocaitions.. not endpoints...

You are always allowed to have an ENDPOINT with a set of
addresses... Associations have to be able to use subsets
of the addresses... period.. otherwise forget IPv6..




> To have multiple logical endpoints here is not a problem. But the
> problem is that they are sharing IP-A and IP-B. This not only goes
> against RFC2960 definition, but also becomes the root cause that opens
> the door for the "corner case" and potential danger of other "duplicate
> association" cases.
> 

There is no corner case...there are incorrect implementations :> but
I don't see a corner case any more..


> Of cause, as you showed below, if a solid implementation is vigorously
> checking and double-checking for dup associations at multiple points
> during the initiation process, you may eventually weed them out. But I
> am not sure about just relying on the implementation to catch dup
> associations.
> 

I think we have a fundamental misconnect here.. you keep talking
of splitting endpoints.. I don't see it. 

R

> -Qiaobing
> 
> >
> > Now what we have forgotten about is what our silly endpoint is going
> > to do.
> >
> > The SCTP stack on Host-Z must look to see if a collision is
> > happening.. this is required in RFC2960. So what happens will depend
> > on where the INIT-ACK is sourced from..
> >
> > If by chance its source is IP-A or IP-B ... guess what? the lookup
> > at Host-Z will return the OLD association which is NOT in a COOKIE-WAIT
> > state.. so what will Host-Z doe? Discard the INIT-ACK... This will
> > happen again (i.e. the INIT-timer will expire on Host-Z and this will
> > then repeat until we fail).... No problem..
> >
> > Now what happens if by chance it IS sourced from IP-C? We find the
> > newly forming association. Now it goes throughs its procedures
> > to attempt to gather all the valuable goodies out of the INIT-ACK.
> > In the process it must load its peers addresses to the association.
> > And here is where it SHOULD detect the other association. I know
> > in our implementation one of the first things I do when loading
> > an address is do a lookup on the address I am loading.. I want
> > to make sure it is in the association only ONCE. In the process
> > I get back the association ... one of the checks I do is verify
> > that it is the same association that I am working on loading
> > address to .. if not.. guess what ? I detect the error and
> > bail out.. dropping the INIT-ACK... so .. same result.. we continue
> > to send out INIT's and drop the result.. thus the IP-C association
> > just never forms..
> >
> > Now is this clear in any document that you must look for addresses
> > as you load them.. no its not specified since it is an implementation
> > detail. I will admit it is an important one.. Should it be documented
> > that you SHOULD attempt to detect this? Maybe but I am not sure
> > where this would fit..
> >
> > But in the end my basic premise is that if the implementation is
> > coded right the silly endpoint will self-correct and NOT bring up
> > this strange association.
> >
> > I don't think we can dis-allow adding an new address... otherwise
> > a true restarting association could not come up if I just added
> > an address until it aborted ... so I don't like that idea...
> >
> > Let me also note .. say my implementation hits a peer that does
> > NOT catch this.. guess what.. the same code is going to run to
> > load the addresses from the cookie.. It too will find this
> > double association setup and drop the cookie...
> >
> > So my contention is that this corner case is NOT an issue if
> > the address loading is done properly by the implementation...
> >
> > R
> >

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Apr 29 22:30:31 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA26843
	for <tsvwg-archive@odin.ietf.org>; Mon, 29 Apr 2002 22:30:19 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id WAA11791
	for tsvwg-archive@odin.ietf.org; Mon, 29 Apr 2002 22:30:22 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id VAA10352;
	Mon, 29 Apr 2002 21:59:46 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id VAA10323
	for <tsvwg@ns.ietf.org>; Mon, 29 Apr 2002 21:59:43 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA26317
	for <tsvwg@ietf.org>; Mon, 29 Apr 2002 21:59:39 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3U1xHH60895;
	Mon, 29 Apr 2002 20:59:18 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCDFA75.E6862A7@stewart.chicago.il.us>
Date: Mon, 29 Apr 2002 20:59:17 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: bidulock@openss7.org
CC: Qiaobing Xie <qxie1@email.mot.com>,
        Michael Tuexen <Michael.Tuexen@icn.siemens.de>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, s_balani@trillium.com,
        mo_khan@trillium.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <875228B9-5BAC-11D6-ACAA-0030654C1AB6@icn.siemens.de> <3CCDDF38.6EA2990E@email.mot.com> <3CCDE3BC.A623629@stewart.chicago.il.us> <20020429195039.J15434@openss7.org>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

"Brian F. G. Bidulock" wrote:
> 
> Randall,
> 
> On Mon, 29 Apr 2002, Randall Stewart wrote:
> 
> [snip]
> >
> > When you attempt to bind IP-Y to the endpoint for port 2222 you must
> > verify that NO OTHER endpoint has it bound.. as long as that is true..
> 
> There is no such restriction in TCP.  It is quite permissible (and
> falls under the category of local port reuse) to have the following
> binding:
> 
Oh yes there is.. try doing this

1) open a socket(tcp)
2) bind IP-A:port 2222 to it
3) open a new socket(tcp)
4) bind IP-A:port 2222 to it.

It will not work.. sorry 


>     E1 [IPA/PortX] ---
>     E2 [IPA/PortX] ---
> 
> In fact these may start off life as:
> 
>     E1 [IPA/Port0] --
>     E2 [IPA/Port0] --
> 
> And when they connect, they autobind with local port reuse and are
> permitted to become:

I am not talking about the ephemeral case here.. Yes I agree
a ENDPOINT will spawn off multiple connected sockets reusing the
port number and the single IP address it has (for TCP that is).

This is not a problem... And I don't see it is a problem with
SCTP either.. An accepted assocaition (on another socket) is
an association under that original listening endpoint. I think
this is what is confusing Qiaobing. I don't see the new socket
as being a seperate endpoint.. it is indirectly associated
with the listener... If another INIT arrives it is NOT
the connected socket that will get the INIT but the 
listner... the listner WILL have to look up and find the connected
socket ... for the collision handling.. but the endpoint that
handles the INIT will be the one listening..

> 
>     E1 [IPA/PortX] ------------------  [IPB/PortY]
>     E2 [IPA/PortX] ------------------  [IPB/PortZ]
> 
> (These ports were never listening.)

By the way I don't see the above two as seperate endpoints. The
endpoint to me is represented by the listener.. these are
not endpoints but associations of that listening endpoint.

> 
> Why does SCTP require that this not happen when it is not only
> permissible but a common port reuse strategy in TCP?
> 

I don't see that it is wrong.. and I think you are misreading
me...

What must be prevented however is
 
E1 [IPA/PortX] ---------A1------------- [IPB/PortZ]
E2 [IPA/PortX] ---------A2------------- [IPB/PortZ]


So:
MachineA has [IPA/IPB]                         Machine B has [IPZ]
E1 [IPA/Port 1000] <---------------A1---------> [IPZ/Port 2000]
and
E2 [IPB/Port 1000] <---------------A2---------> [IPZ/Port 2000]

Now if I add IP-C to machine A

Only one of these guys can have IP-C added...

E1 or E2....<the listner and its spawned associations>

Pick whichever one you want..

Otherwise you end up with

E1 [IPA,IPC/Port 1000]<--------------A1----->[IPZ/Port 2000]
and
E2 [IPA,IPC/Port 1000]<--------------A2----->[IPZ/Port 2000]

and when a packet comes along to the kernel of machine a:
           <----[To:IPC port 1000 From IPZ port 2000]----
or
        ----[From IPC/Port 1000 To IPZ port 2000]--------->

You have NO way to distingush which of the associations the
packets belong to..


I guess it all comes down to the definition of endpoint:

To me an Endpoint is either the connect() or listening() socket
and any associations it spawns. In another view the one that
binds the port. 

If I just do a implicit socket/connect for example.. in the process
of doing the connect an implicit bind will happen of a ephemeral
port and at that point that socket becomes an SCTP endpoint
bound to that port. Any associations it spawns under that
by doing connect() or closing it and changing to a listen
are part of that endpoint... 

If I do a socket/bind/listen or a socket/<bind>/connect this
makes me an endpoint.

For the UDP model I do a socket/bind/listen as well or a 
socket/connect... same result... the socket becomes an endpoint.

At that point associations underneath the endpoint (ones
created by implicit connection setup/ connect() calls/ 
accept() calls or even sctp_peel_offs are STILL part of that
endpoint.. and can use all or a sub-set of the addresses
of the endpoint. When we add an address to a endpoint it
must be to the base endpoint... i.e. the one that
has the binding. This then authorizes all of the sub-servant 
associations of the endpoint to also do an ADD-IP if they wish.

Thats how I see it anyway...

R





> --brian
> 
> > No problem... You add that one to the endpoint... and if some
> > association can only use a subset of the addresses so be it...
> >
> > Now as to the corner case.. I think I have shown that this is NOT
> > a problem... proper coding will take care of this issue..
> >
> > Do we need some documentation somewhere? Maybe...
> >
> > R
> > --
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> --
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 30 01:24:07 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA00287
	for <tsvwg-archive@odin.ietf.org>; Tue, 30 Apr 2002 01:24:07 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id BAA28201
	for tsvwg-archive@odin.ietf.org; Tue, 30 Apr 2002 01:24:09 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id AAA18455;
	Tue, 30 Apr 2002 00:49:11 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id AAA18426
	for <tsvwg@optimus.ietf.org>; Tue, 30 Apr 2002 00:49:08 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-230.dallas.net [209.44.42.230])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA29867
	for <tsvwg@ietf.org>; Tue, 30 Apr 2002 00:49:02 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id XAA02426;
	Mon, 29 Apr 2002 23:48:27 -0500
Date: Mon, 29 Apr 2002 23:48:27 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Randall Stewart <randall@stewart.chicago.il.us>
Cc: Qiaobing Xie <qxie1@email.mot.com>,
        Michael Tuexen <Michael.Tuexen@icn.siemens.de>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, s_balani@trillium.com,
        mo_khan@trillium.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
Message-ID: <20020429234827.C32335@openss7.org>
Reply-To: bidulock@openss7.org
References: <875228B9-5BAC-11D6-ACAA-0030654C1AB6@icn.siemens.de> <3CCDDF38.6EA2990E@email.mot.com> <3CCDE3BC.A623629@stewart.chicago.il.us> <20020429195039.J15434@openss7.org> <3CCDFA75.E6862A7@stewart.chicago.il.us>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <3CCDFA75.E6862A7@stewart.chicago.il.us>; from randall@stewart.chicago.il.us on Mon, Apr 29, 2002 at 08:59:17PM -0500
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Randall,

Please see comments below...

On Mon, 29 Apr 2002, Randall Stewart wrote:

> "Brian F. G. Bidulock" wrote:
> > 
> > Randall,
> > 
> > On Mon, 29 Apr 2002, Randall Stewart wrote:
> > 
> > [snip]
> > >
> > > When you attempt to bind IP-Y to the endpoint for port 2222 you must
> > > verify that NO OTHER endpoint has it bound.. as long as that is true..
> > 
> > There is no such restriction in TCP.  It is quite permissible (and
> > falls under the category of local port reuse) to have the following
> > binding:
> > 
> Oh yes there is.. try doing this
> 
> 1) open a socket(tcp)
> 2) bind IP-A:port 2222 to it
> 3) open a new socket(tcp)
> 4) bind IP-A:port 2222 to it.
> 
> It will not work.. sorry 

Err, yes it does, if you set a standard socket option...

Compile and run the following:

#include <sys/types.h>
#include <sys/socket.h>
#include <stdio.h>
#include <errno.h>
#include <netinet/in.h>
#include <netdb.h>
#include <string.h>

main()
{
	int s1, s2, flag = 1;
	struct sockaddr_in addr1 = { AF_INET, 0, { 0 } };
	struct hostent *haddr;
	haddr = gethostbyname("127.0.0.1");
	addr1.sin_port = htons(10000);
	addr1.sin_addr.s_addr = *(uint32_t *)haddr->h_addr;
	if ((s1 = socket(PF_INET, SOCK_STREAM, IPPROTO_TCP)) < 0 ) {
		perror("test);
		exit(2);
	}
	if ((s2 = socket(PF_INET, SOCK_STREAM, IPPROTO_TCP)) < 0 ) {
		perror("test);
		exit(2);
	}
	if (setsockopt(s1, SOL_SOCKET, SO_REUSEADDR, &flag, sizeof(flag)) < 0 ) {
		perror("test);
		exit(2);
	}
	if (setsockopt(s2, SOL_SOCKET, SO_REUSEADDR, &flag, sizeof(flag)) < 0 ) {
		perror("test);
		exit(2);
	}
	if (bind(s1, (struct sockaddr *)&addr1, sizeof(addr1)) < 0) {
		perror("test);
		exit(2);
	}
	if (bind(s2, (struct sockaddr *)&addr1, sizeof(addr1)) < 0) {
		perror("test);
		exit(2);
	}
	printf("Q.E.D.\n");
}

Output:

#> gcc -o test test.c

#> ./test
Q.E.D.

Now, is there any reason why I should not be able to do the same
for SCTP?  Also, why should the system stop me from overlapping
address lists with the same port number?  TCP doesn't stop me...

It is only once the address is fully formed that this overlap is
checked for TCP.  If you attempt to connect the second socket to
the same destination transport address (IP and port) as the first
one has connected, it will give you an error.  Again, why should
SCTP be any different?

Therefore, it is quite rational to have for SCTP:

   E1 [IPB, IPA port 10000] ---
   E2 [IPC, IPA port 10000] ---

regardless of the statements on Page 13 of the RFC.  The "MUST NOT"
still stands in that these two endpoints cannot connect to the same
destination transport address because they share IPA port 10000.
But there is nothing to stop me from binding and defining an endpoint
in this way except a rather off-base statement that does not consider
the longstanding use of SO_REUSEADDR.

--brian

> 
> 
> >     E1 [IPA/PortX] ---
> >     E2 [IPA/PortX] ---
> > 
> > In fact these may start off life as:
> > 
> >     E1 [IPA/Port0] --
> >     E2 [IPA/Port0] --
> > 
> > And when they connect, they autobind with local port reuse and are
> > permitted to become:
> 
> I am not talking about the ephemeral case here.. Yes I agree
> a ENDPOINT will spawn off multiple connected sockets reusing the
> port number and the single IP address it has (for TCP that is).
> 
> This is not a problem... And I don't see it is a problem with
> SCTP either.. An accepted assocaition (on another socket) is
> an association under that original listening endpoint. I think
> this is what is confusing Qiaobing. I don't see the new socket
> as being a seperate endpoint.. it is indirectly associated
> with the listener... If another INIT arrives it is NOT
> the connected socket that will get the INIT but the 
> listner... the listner WILL have to look up and find the connected
> socket ... for the collision handling.. but the endpoint that
> handles the INIT will be the one listening..
> 
> > 
> >     E1 [IPA/PortX] ------------------  [IPB/PortY]
> >     E2 [IPA/PortX] ------------------  [IPB/PortZ]
> > 
> > (These ports were never listening.)
> 
> By the way I don't see the above two as seperate endpoints. The
> endpoint to me is represented by the listener.. these are
> not endpoints but associations of that listening endpoint.
> 
> > 
> > Why does SCTP require that this not happen when it is not only
> > permissible but a common port reuse strategy in TCP?
> > 
> 
> I don't see that it is wrong.. and I think you are misreading
> me...
> 
> What must be prevented however is
>  
> E1 [IPA/PortX] ---------A1------------- [IPB/PortZ]
> E2 [IPA/PortX] ---------A2------------- [IPB/PortZ]
> 
> 
> So:
> MachineA has [IPA/IPB]                         Machine B has [IPZ]
> E1 [IPA/Port 1000] <---------------A1---------> [IPZ/Port 2000]
> and
> E2 [IPB/Port 1000] <---------------A2---------> [IPZ/Port 2000]
> 
> Now if I add IP-C to machine A
> 
> Only one of these guys can have IP-C added...
> 
> E1 or E2....<the listner and its spawned associations>
> 
> Pick whichever one you want..
> 
> Otherwise you end up with
> 
> E1 [IPA,IPC/Port 1000]<--------------A1----->[IPZ/Port 2000]
> and
> E2 [IPA,IPC/Port 1000]<--------------A2----->[IPZ/Port 2000]
> 
> and when a packet comes along to the kernel of machine a:
>            <----[To:IPC port 1000 From IPZ port 2000]----
> or
>         ----[From IPC/Port 1000 To IPZ port 2000]--------->
> 
> You have NO way to distingush which of the associations the
> packets belong to..
> 
> 
> I guess it all comes down to the definition of endpoint:
> 
> To me an Endpoint is either the connect() or listening() socket
> and any associations it spawns. In another view the one that
> binds the port. 
> 
> If I just do a implicit socket/connect for example.. in the process
> of doing the connect an implicit bind will happen of a ephemeral
> port and at that point that socket becomes an SCTP endpoint
> bound to that port. Any associations it spawns under that
> by doing connect() or closing it and changing to a listen
> are part of that endpoint... 
> 
> If I do a socket/bind/listen or a socket/<bind>/connect this
> makes me an endpoint.
> 
> For the UDP model I do a socket/bind/listen as well or a 
> socket/connect... same result... the socket becomes an endpoint.
> 
> At that point associations underneath the endpoint (ones
> created by implicit connection setup/ connect() calls/ 
> accept() calls or even sctp_peel_offs are STILL part of that
> endpoint.. and can use all or a sub-set of the addresses
> of the endpoint. When we add an address to a endpoint it
> must be to the base endpoint... i.e. the one that
> has the binding. This then authorizes all of the sub-servant 
> associations of the endpoint to also do an ADD-IP if they wish.
> 
> Thats how I see it anyway...
> 
> R
> 
> 
> 
> 
> 
> > --brian
> > 
> > > No problem... You add that one to the endpoint... and if some
> > > association can only use a subset of the addresses so be it...
> > >
> > > Now as to the corner case.. I think I have shown that this is NOT
> > > a problem... proper coding will take care of this issue..
> > >
> > > Do we need some documentation somewhere? Maybe...
> > >
> > > R
> > > --
> > > Randall R. Stewart
> > > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> > >
> > > _______________________________________________
> > > tsvwg mailing list
> > > tsvwg@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > 
> > --
> > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> >                         ¦ Therefore  all  progress  depends on the ¦
> >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 30 08:36:01 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA16876
	for <tsvwg-archive@odin.ietf.org>; Tue, 30 Apr 2002 08:35:48 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA20474
	for tsvwg-archive@odin.ietf.org; Tue, 30 Apr 2002 08:35:50 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA18248;
	Tue, 30 Apr 2002 08:08:17 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA18219
	for <tsvwg@optimus.ietf.org>; Tue, 30 Apr 2002 08:08:15 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA15217
	for <tsvwg@ietf.org>; Tue, 30 Apr 2002 08:08:11 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3UC7FH61932;
	Tue, 30 Apr 2002 07:07:16 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCE88F3.16AAA93E@stewart.chicago.il.us>
Date: Tue, 30 Apr 2002 07:07:15 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: bidulock@openss7.org
CC: Qiaobing Xie <qxie1@email.mot.com>,
        Michael Tuexen <Michael.Tuexen@icn.siemens.de>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, s_balani@trillium.com,
        mo_khan@trillium.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <875228B9-5BAC-11D6-ACAA-0030654C1AB6@icn.siemens.de> <3CCDDF38.6EA2990E@email.mot.com> <3CCDE3BC.A623629@stewart.chicago.il.us> <20020429195039.J15434@openss7.org> <3CCDFA75.E6862A7@stewart.chicago.il.us> <20020429234827.C32335@openss7.org>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Brian:

A comment for you and a question please see below:


"Brian F. G. Bidulock" wrote:
> 
> Randall,
> 
> Please see comments below...
> 
> On Mon, 29 Apr 2002, Randall Stewart wrote:
> 
> > "Brian F. G. Bidulock" wrote:
> > >
> > > Randall,
> > >
> > > On Mon, 29 Apr 2002, Randall Stewart wrote:
> > >
> > > [snip]
> > > >
> > > > When you attempt to bind IP-Y to the endpoint for port 2222 you must
> > > > verify that NO OTHER endpoint has it bound.. as long as that is true..
> > >
> > > There is no such restriction in TCP.  It is quite permissible (and
> > > falls under the category of local port reuse) to have the following
> > > binding:
> > >
> > Oh yes there is.. try doing this
> >
> > 1) open a socket(tcp)
> > 2) bind IP-A:port 2222 to it
> > 3) open a new socket(tcp)
> > 4) bind IP-A:port 2222 to it.
> >
> > It will not work.. sorry
> 
> Err, yes it does, if you set a standard socket option...
> 

....Code deleted....

> It is only once the address is fully formed that this overlap is
> checked for TCP.  If you attempt to connect the second socket to
> the same destination transport address (IP and port) as the first
> one has connected, it will give you an error.  Again, why should
> SCTP be any different?
> 
> Therefore, it is quite rational to have for SCTP:
> 
>    E1 [IPB, IPA port 10000] ---
>    E2 [IPC, IPA port 10000] ---
> 
> regardless of the statements on Page 13 of the RFC.  The "MUST NOT"
> still stands in that these two endpoints cannot connect to the same
> destination transport address because they share IPA port 10000.
> But there is nothing to stop me from binding and defining an endpoint
> in this way except a rather off-base statement that does not consider
> the longstanding use of SO_REUSEADDR.
> 

Note I did NOT say I added any special socket option... If you
take out the additional option it will NOT suceed.

The reason this CAN work for TCP is that the peer connecting
is a single endpoint... Great care must be taken when using
this socket option... that is why it is an additional option
and NOT default behavior. TCP would get in the same trouble
if both sides of a connection did the above and both sides
connected to each other... you would end up with

[IP-A Port 1000]------TCP connection 1------[IP-B Port 2000]
[IP-A Port 1000]------TCP connection 2------[IP-B Port 2000]

Which is a fundamental violation of the definition of a TCP
connection. I bet this is why it is an option.. you have to
specifically set this up in order to have it occur (and
frankly I think it would confuse most stacks.. data would
not go to the correct places if things would function at all).
The base socket system if used without auxillary options will
NOT allow you to do the above...  Going back to my early
email..

...But fundamentally you have NOT answered my question below.
How will you tell which endpoint gets which packet? 

Please answer the question below...

> >
> > I am not talking about the ephemeral case here.. Yes I agree
> > a ENDPOINT will spawn off multiple connected sockets reusing the
> > port number and the single IP address it has (for TCP that is).
> >
> > This is not a problem... And I don't see it is a problem with
> > SCTP either.. An accepted assocaition (on another socket) is
> > an association under that original listening endpoint. I think
> > this is what is confusing Qiaobing. I don't see the new socket
> > as being a seperate endpoint.. it is indirectly associated
> > with the listener... If another INIT arrives it is NOT
> > the connected socket that will get the INIT but the
> > listner... the listner WILL have to look up and find the connected
> > socket ... for the collision handling.. but the endpoint that
> > handles the INIT will be the one listening..
> >
> > >
> > >     E1 [IPA/PortX] ------------------  [IPB/PortY]
> > >     E2 [IPA/PortX] ------------------  [IPB/PortZ]
> > >
> > > (These ports were never listening.)
> >
> > By the way I don't see the above two as seperate endpoints. The
> > endpoint to me is represented by the listener.. these are
> > not endpoints but associations of that listening endpoint.
> >
> > >
> > > Why does SCTP require that this not happen when it is not only
> > > permissible but a common port reuse strategy in TCP?
> > >
> >
> > I don't see that it is wrong.. and I think you are misreading
> > me...
> >
> > What must be prevented however is
> >
> > E1 [IPA/PortX] ---------A1------------- [IPB/PortZ]
> > E2 [IPA/PortX] ---------A2------------- [IPB/PortZ]
> >
> >
> > So:
> > MachineA has [IPA/IPB]                         Machine B has [IPZ]
> > E1 [IPA/Port 1000] <---------------A1---------> [IPZ/Port 2000]
> > and
> > E2 [IPB/Port 1000] <---------------A2---------> [IPZ/Port 2000]
> >
> > Now if I add IP-C to machine A
> >
> > Only one of these guys can have IP-C added...
> >
> > E1 or E2....<the listner and its spawned associations>
> >
> > Pick whichever one you want..
> >
> > Otherwise you end up with
> >
> > E1 [IPA,IPC/Port 1000]<--------------A1----->[IPZ/Port 2000]
> > and
> > E2 [IPA,IPC/Port 1000]<--------------A2----->[IPZ/Port 2000]
> >
> > and when a packet comes along to the kernel of machine a:
> >            <----[To:IPC port 1000 From IPZ port 2000]----
> > or
> >         ----[From IPC/Port 1000 To IPZ port 2000]--------->
> >
> > You have NO way to distingush which of the associations the
> > packets belong to..
> >
> >
> > I guess it all comes down to the definition of endpoint:
> >
> > To me an Endpoint is either the connect() or listening() socket
> > and any associations it spawns. In another view the one that
> > binds the port.
> >
> > If I just do a implicit socket/connect for example.. in the process
> > of doing the connect an implicit bind will happen of a ephemeral
> > port and at that point that socket becomes an SCTP endpoint
> > bound to that port. Any associations it spawns under that
> > by doing connect() or closing it and changing to a listen
> > are part of that endpoint...
> >
> > If I do a socket/bind/listen or a socket/<bind>/connect this
> > makes me an endpoint.
> >
> > For the UDP model I do a socket/bind/listen as well or a
> > socket/connect... same result... the socket becomes an endpoint.
> >
> > At that point associations underneath the endpoint (ones
> > created by implicit connection setup/ connect() calls/
> > accept() calls or even sctp_peel_offs are STILL part of that
> > endpoint.. and can use all or a sub-set of the addresses
> > of the endpoint. When we add an address to a endpoint it
> > must be to the base endpoint... i.e. the one that
> > has the binding. This then authorizes all of the sub-servant
> > associations of the endpoint to also do an ADD-IP if they wish.
> >
> > Thats how I see it anyway...
> >
> > R
> >
> >
> >
> >
> >
> > > --brian
> > >
> > > > No problem... You add that one to the endpoint... and if some
> > > > association can only use a subset of the addresses so be it...
> > > >
> > > > Now as to the corner case.. I think I have shown that this is NOT
> > > > a problem... proper coding will take care of this issue..
> > > >
> > > > Do we need some documentation somewhere? Maybe...
> > > >
> > > > R
> > > > --
> > > > Randall R. Stewart
> > > > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> > > >
> > > > _______________________________________________
> > > > tsvwg mailing list
> > > > tsvwg@ietf.org
> > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > >
> > > --
> > > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> > >                         ¦ Therefore  all  progress  depends on the ¦
> > >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> >
> > --
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 
> --
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 30 14:42:11 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA03680
	for <tsvwg-archive@odin.ietf.org>; Tue, 30 Apr 2002 14:42:11 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA17109
	for tsvwg-archive@odin.ietf.org; Tue, 30 Apr 2002 14:42:14 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA15216;
	Tue, 30 Apr 2002 14:17:07 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA15182
	for <tsvwg@optimus.ietf.org>; Tue, 30 Apr 2002 14:17:05 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-230.dallas.net [209.44.42.230])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA02362
	for <tsvwg@ietf.org>; Tue, 30 Apr 2002 14:16:59 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id NAA13000;
	Tue, 30 Apr 2002 13:16:18 -0500
Date: Tue, 30 Apr 2002 13:16:18 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Randall Stewart <randall@stewart.chicago.il.us>
Cc: Qiaobing Xie <qxie1@email.mot.com>,
        Michael Tuexen <Michael.Tuexen@icn.siemens.de>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, s_balani@trillium.com,
        mo_khan@trillium.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
Message-ID: <20020430131618.A2547@openss7.org>
Reply-To: bidulock@openss7.org
References: <875228B9-5BAC-11D6-ACAA-0030654C1AB6@icn.siemens.de> <3CCDDF38.6EA2990E@email.mot.com> <3CCDE3BC.A623629@stewart.chicago.il.us> <20020429195039.J15434@openss7.org> <3CCDFA75.E6862A7@stewart.chicago.il.us> <20020429234827.C32335@openss7.org> <3CCE88F3.16AAA93E@stewart.chicago.il.us>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <3CCE88F3.16AAA93E@stewart.chicago.il.us>; from randall@stewart.chicago.il.us on Tue, Apr 30, 2002 at 07:07:15AM -0500
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Randall,

On Tue, 30 Apr 2002, Randall Stewart wrote:

> 
> Note I did NOT say I added any special socket option... If you
> take out the additional option it will NOT suceed.
> 
> The reason this CAN work for TCP is that the peer connecting
> is a single endpoint... Great care must be taken when using
> this socket option... that is why it is an additional option
> and NOT default behavior. TCP would get in the same trouble
> if both sides of a connection did the above and both sides
> connected to each other... you would end up with
> 
> [IP-A Port 1000]------TCP connection 1------[IP-B Port 2000]
> [IP-A Port 1000]------TCP connection 2------[IP-B Port 2000]

The standard error EADDRINUSE would be returned by
second connect() call.   It takes no big effort: in
a NET4 implementation a check is made that the connection
does not already exist in the established hashes.

> 
> Which is a fundamental violation of the definition of a TCP
> connection. I bet this is why it is an option.. you have to
> specifically set this up in order to have it occur (and
> frankly I think it would confuse most stacks.. data would
> not go to the correct places if things would function at all).
> The base socket system if used without auxillary options will
> NOT allow you to do the above...  Going back to my early
> email..
> 
> ...But fundamentally you have NOT answered my question below.
> How will you tell which endpoint gets which packet? 

These endpoints are not listening.  When they engage in fully
formed connections the endpoint is determined from the established
hashes which requires a match on vtag, source address, source port,
destination address and destination port.  As long as this set
uniquely identifies an endpoint (and listen() and connect() will
make sure that this is true) there is never any problem, just
as there is never any problem for TCP above.

--brian

> 
> Please answer the question below...
> 
> > >
> > > I am not talking about the ephemeral case here.. Yes I agree
> > > a ENDPOINT will spawn off multiple connected sockets reusing the
> > > port number and the single IP address it has (for TCP that is).
> > >
> > > This is not a problem... And I don't see it is a problem with
> > > SCTP either.. An accepted assocaition (on another socket) is
> > > an association under that original listening endpoint. I think
> > > this is what is confusing Qiaobing. I don't see the new socket
> > > as being a seperate endpoint.. it is indirectly associated
> > > with the listener... If another INIT arrives it is NOT
> > > the connected socket that will get the INIT but the
> > > listner... the listner WILL have to look up and find the connected
> > > socket ... for the collision handling.. but the endpoint that
> > > handles the INIT will be the one listening..
> > >
> > > >
> > > >     E1 [IPA/PortX] ------------------  [IPB/PortY]
> > > >     E2 [IPA/PortX] ------------------  [IPB/PortZ]
> > > >
> > > > (These ports were never listening.)
> > >
> > > By the way I don't see the above two as seperate endpoints. The
> > > endpoint to me is represented by the listener.. these are
> > > not endpoints but associations of that listening endpoint.
> > >
> > > >
> > > > Why does SCTP require that this not happen when it is not only
> > > > permissible but a common port reuse strategy in TCP?
> > > >
> > >
> > > I don't see that it is wrong.. and I think you are misreading
> > > me...
> > >
> > > What must be prevented however is
> > >
> > > E1 [IPA/PortX] ---------A1------------- [IPB/PortZ]
> > > E2 [IPA/PortX] ---------A2------------- [IPB/PortZ]
> > >
> > >
> > > So:
> > > MachineA has [IPA/IPB]                         Machine B has [IPZ]
> > > E1 [IPA/Port 1000] <---------------A1---------> [IPZ/Port 2000]
> > > and
> > > E2 [IPB/Port 1000] <---------------A2---------> [IPZ/Port 2000]
> > >
> > > Now if I add IP-C to machine A
> > >
> > > Only one of these guys can have IP-C added...
> > >
> > > E1 or E2....<the listner and its spawned associations>
> > >
> > > Pick whichever one you want..
> > >
> > > Otherwise you end up with
> > >
> > > E1 [IPA,IPC/Port 1000]<--------------A1----->[IPZ/Port 2000]
> > > and
> > > E2 [IPA,IPC/Port 1000]<--------------A2----->[IPZ/Port 2000]
> > >
> > > and when a packet comes along to the kernel of machine a:
> > >            <----[To:IPC port 1000 From IPZ port 2000]----
> > > or
> > >         ----[From IPC/Port 1000 To IPZ port 2000]--------->
> > >
> > > You have NO way to distingush which of the associations the
> > > packets belong to..
> > >
> > >
> > > I guess it all comes down to the definition of endpoint:
> > >
> > > To me an Endpoint is either the connect() or listening() socket
> > > and any associations it spawns. In another view the one that
> > > binds the port.
> > >
> > > If I just do a implicit socket/connect for example.. in the process
> > > of doing the connect an implicit bind will happen of a ephemeral
> > > port and at that point that socket becomes an SCTP endpoint
> > > bound to that port. Any associations it spawns under that
> > > by doing connect() or closing it and changing to a listen
> > > are part of that endpoint...
> > >
> > > If I do a socket/bind/listen or a socket/<bind>/connect this
> > > makes me an endpoint.
> > >
> > > For the UDP model I do a socket/bind/listen as well or a
> > > socket/connect... same result... the socket becomes an endpoint.
> > >
> > > At that point associations underneath the endpoint (ones
> > > created by implicit connection setup/ connect() calls/
> > > accept() calls or even sctp_peel_offs are STILL part of that
> > > endpoint.. and can use all or a sub-set of the addresses
> > > of the endpoint. When we add an address to a endpoint it
> > > must be to the base endpoint... i.e. the one that
> > > has the binding. This then authorizes all of the sub-servant
> > > associations of the endpoint to also do an ADD-IP if they wish.
> > >
> > > Thats how I see it anyway...
> > >
> > > R
> > >
> > >
> > >
> > >
> > >
> > > > --brian
> > > >
> > > > > No problem... You add that one to the endpoint... and if some
> > > > > association can only use a subset of the addresses so be it...
> > > > >
> > > > > Now as to the corner case.. I think I have shown that this is NOT
> > > > > a problem... proper coding will take care of this issue..
> > > > >
> > > > > Do we need some documentation somewhere? Maybe...
> > > > >
> > > > > R
> > > > > --
> > > > > Randall R. Stewart
> > > > > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> > > > >
> > > > > _______________________________________________
> > > > > tsvwg mailing list
> > > > > tsvwg@ietf.org
> > > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > > >
> > > > --
> > > > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > > > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > > > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> > > >                         ¦ Therefore  all  progress  depends on the ¦
> > > >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> > >
> > > --
> > > Randall R. Stewart
> > > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> > 
> > --
> > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> >                         ¦ Therefore  all  progress  depends on the ¦
> >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> > 
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 30 14:55:16 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA04244
	for <tsvwg-archive@odin.ietf.org>; Tue, 30 Apr 2002 14:55:16 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA17460
	for tsvwg-archive@odin.ietf.org; Tue, 30 Apr 2002 14:55:19 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA15719;
	Tue, 30 Apr 2002 14:27:21 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA15688
	for <tsvwg@optimus.ietf.org>; Tue, 30 Apr 2002 14:27:18 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA02812
	for <tsvwg@ietf.org>; Tue, 30 Apr 2002 14:27:14 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3UIQoH62739;
	Tue, 30 Apr 2002 13:26:50 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCEE1E9.4361187B@stewart.chicago.il.us>
Date: Tue, 30 Apr 2002 13:26:50 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: bidulock@openss7.org
CC: Qiaobing Xie <qxie1@email.mot.com>,
        Michael Tuexen <Michael.Tuexen@icn.siemens.de>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, s_balani@trillium.com,
        mo_khan@trillium.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <875228B9-5BAC-11D6-ACAA-0030654C1AB6@icn.siemens.de> <3CCDDF38.6EA2990E@email.mot.com> <3CCDE3BC.A623629@stewart.chicago.il.us> <20020429195039.J15434@openss7.org> <3CCDFA75.E6862A7@stewart.chicago.il.us> <20020429234827.C32335@openss7.org> <3CCE88F3.16AAA93E@stewart.chicago.il.us> <20020430131618.A2547@openss7.org>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

"Brian F. G. Bidulock" wrote:

> These endpoints are not listening.  When they engage in fully
> formed connections the endpoint is determined from the established
> hashes which requires a match on vtag, source address, source port,
> destination address and destination port.  As long as this set
> uniquely identifies an endpoint (and listen() and connect() will
> make sure that this is true) there is never any problem, just
> as there is never any problem for TCP above.
> 

Vtags have never been defined as part of the identification
of an SCTP association. They are used in place of the
pseudo header and as a defense against blind attacks.. not
as part of association discrimination... Sorry this won't work.
Two associations are not allowed in this case.. just like
your TCP description would return a INUSE error...

One can easily prevent the TCP case.. i.e. at the connect
check the hash.... I guess the ADD-IP case also can
be prevented when one tries to add the IP-C to the second
association you would simply get the same error :>

Thinking about it.. we do this now.. since when you attempt
the sctp_bindx you would get such an error :>

R


-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 30 16:11:05 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA07202
	for <tsvwg-archive@odin.ietf.org>; Tue, 30 Apr 2002 16:11:05 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA22533
	for tsvwg-archive@odin.ietf.org; Tue, 30 Apr 2002 16:11:08 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA20729;
	Tue, 30 Apr 2002 15:46:41 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA15104
	for <tsvwg@optimus.ietf.org>; Tue, 30 Apr 2002 07:16:26 -0400 (EDT)
Received: from CNRI.Reston.VA.US (localhost [127.0.0.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA13246;
	Tue, 30 Apr 2002 07:16:19 -0400 (EDT)
Message-Id: <200204301116.HAA13246@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: IETF-Announce: ;
Cc: tsvwg@ietf.org
From: Internet-Drafts@ietf.org
Reply-to: Internet-Drafts@ietf.org
Date: Tue, 30 Apr 2002 07:16:18 -0400
Subject: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-tcp-nonce-03.txt
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the Transport Area Working Group Working Group of the IETF.

	Title		: Robust ECN Signaling with Nonces
	Author(s)	: D. Wetherall, N. Spring, D. Ely
	Filename	: draft-ietf-tsvwg-tcp-nonce-03.txt
	Pages		: 12
	Date		: 29-Apr-02
	
This note describes the ECN-nonce, an optional addition to ECN that
protects against accidental or malicious concealment of marked
packets from the TCP sender.  It improves the robustness of
congestion control by preventing receivers from exploiting ECN to
gain an unfair share of network bandwidth.  The ECN-nonce uses the
two ECT codepoints in the ECN field of the IP header, and requires a
flag in the TCP header.  It is computationally efficient for both
routers and hosts.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-tcp-nonce-03.txt

To remove yourself from the IETF Announcement list, send a message to 
ietf-announce-request with the word unsubscribe in the body of the message.

Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-tsvwg-tcp-nonce-03.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-tsvwg-tcp-nonce-03.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<20020429151450.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-tsvwg-tcp-nonce-03.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-tsvwg-tcp-nonce-03.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<20020429151450.I-D@ietf.org>

--OtherAccess--

--NextPart--




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 30 16:12:45 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA07255
	for <tsvwg-archive@odin.ietf.org>; Tue, 30 Apr 2002 16:12:45 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA22637
	for tsvwg-archive@odin.ietf.org; Tue, 30 Apr 2002 16:12:48 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA20889;
	Tue, 30 Apr 2002 15:47:38 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA25124
	for <tsvwg@optimus.ietf.org>; Tue, 30 Apr 2002 09:52:41 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA20392
	for <tsvwg@ietf.org>; Tue, 30 Apr 2002 09:52:37 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3UDq5H62186;
	Tue, 30 Apr 2002 08:52:05 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCEA184.1D1E29FD@stewart.chicago.il.us>
Date: Tue, 30 Apr 2002 08:52:05 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Jon Grimm <jgrimm@austin.ibm.com>
CC: Qiaobing Xie <qxie1@email.mot.com>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, bidulock@openss7.org,
        Michael.Tuexen@icn.siemens.de, s_balani@trillium.com,
        mo_khan@trillium.com, h_pillai@trillium.com, jheitz@lucent.com,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFB4@esebe017.NOE.Nokia.com> <3CCD844A.DB1BC960@austin.ibm.com> <3CCD9267.6456A90E@stewart.chicago.il.us> <3CCDA2AD.CB11491@austin.ibm.com> <3CCDB210.E8FA4663@stewart.chicago.il.us> <3CCDD167.8A0DBC43@email.mot.com> <3CCDDF18.9DDD2030@stewart.chicago.il.us> <3CCDED1D.EFC1E6F0@email.mot.com> <3CCDF641.56A4C9AD@stewart.chicago.il.us> <3CCE9A9F.5010703@austin.ibm.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Jon Grimm wrote:
> 
> Randall Stewart wrote:
> 
> > Qiaobing Xie wrote:
> >
> >>R,
> >>
> >>A couple of comments...
> >>
> >>Randall Stewart wrote:
> >>...
> >>
> >>>Now I have done a bit more thinking on this. The corner case that
> >>>we were worried about was this (if I remember right):
> >>>
> >>>Host A                                          Host Z
> >>>IP-A/IP-B                                       IP-Z
> >>>(An UDP model Endpoint
> >>> Listening on Port 2222)
> >>>        IP-A:Port-2222<-----------------INIT(From IP-Z:Port 3333)
> >>>        ----------INIT-ACK(IP-A,IP-B:Port 2222)-------->
> >>>        <--------------COOKIE-ECHO----------------------
> >>>        ----------------COOKIE-ACK---------------------->
> >>>
> >>>            Now at this point A-1 is up.
> >>>
> >>>Now along the line we do
> >>>
> >>>ifconfg New-Intf IP-C
> >>>
> >>>and magically do
> >>>
> >>>--------------ASCONF(ADD-IP:IP-C)------------------------>
> >>>    <---------------------OperationalError(Unknown Chunk ASCONF)-----
> >>>
> >>>So here we have a potential problem...
> >>>
> >>At this point, would our local stack (I am avoiding to use endpoint
> >>here) actually start to use IP-C as the source addr in any outbound
> >>packets in the current association? If so, that would trigger an abort
> >>from the peer and kill the current association. If not, then strictly
> >>speaking, IP-C has not yet been added to the local **endpoint** that the
> >>current associaiton is using. It has only been made available to the
> >>local stack for future use.
> >>
> >>
> >>>                    now our silly peer does
> >>>
> >>> <------------------------IP-C:2222 INIT(IP-Z:Port 3333)
> >>>
> >>>What do we do?
> >>>
> >>> --------INIT-ACK(IP-A,IP-B,IP-C: Port 2222)---------->IP-Z:Port
> >>>3333--->
> >>>
> >>When the local stack does this, it starts to manage two logically
> >>different local endpoints - the one used by the old association that has
> >>only IP-A and IP-B; the other is intended for the new association, as
> >>advertised in this INIT-ACK, that has IP-A, IP-B, and IP-C.
> >>
> >>
> >
> > No it is managing two different assocaitions.. not endpoints...
> >
> > You are always allowed to have an ENDPOINT with a set of
> > addresses... Associations have to be able to use subsets
> > of the addresses... period.. otherwise forget IPv6..
> >
> >
> 
> I can't see if it matters or not, but how about the delete-ip case?  If
> an endpoint has multiple associations and a subset of the associations
> support delete-ip.   What should happens to the endpoint?  Is the
> address really removed from the endpoint?  There are associations still
> using that address removed from some of the addresses.  I can see where
> I might want to keep the address in the endpoint until all users of the
> address have either ACKed the delete or have terminated.  In this case,
> I wouldn't want to allow a new endpoint to form with that deleted
> address.  Is this warped thinking?
> 
Jon:

well let see... I think the way we handled this is we DID
remove it from the "endpoint" or listener... that way you
will NOT accept new INIT's to this.. of course if the 
address is truely removed from the system its a MOOT point :>

Now each association has its own list of addresses that are
in this "present and not deleted state" or "added but not completed"
state.
That way if somehow a packet came in to that address we would
locate the right association.

I think the big thing here that I have not stated is I don't
believe we allow you to add/delete an IP address to a association.

You do a sctp_bindx() against an endpoint yes.

You have flags that say (default is on in most cases) DO ASCONF
if you can.. So you get this behavior on all your fanned out
associations.. but if I am adding an address and some association
under my endpoint can't.. no big deal it just goes on a negative
list saying.. don't use this address. (And we are in the kernel
so we DO control the IP address that goes on the wire... if you
can't control this address you are probably in rough shape and
unable to do add-ip .. the more I think about it... of course
if you can't control your IP address your stack is kinda broken :>)

But you CAN'T say ... add/delete this address to association X under
this endpoint.

Now for the "connected/peeled-off" socket you can do this .. since doing 
a sctp_bindx() against the connected socket WOULD add it to the
association.. but this of course does NOT change the listening
endpoint... so it has no ramifications against future INIT's that
arrive at the other endpoint that did NOT do a sctp_bindx() on the
address.


R


> 
> <snipped everything else>

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 30 16:13:02 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA07280
	for <tsvwg-archive@odin.ietf.org>; Tue, 30 Apr 2002 16:13:02 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA22666
	for tsvwg-archive@odin.ietf.org; Tue, 30 Apr 2002 16:13:05 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA20993;
	Tue, 30 Apr 2002 15:47:54 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA01588
	for <tsvwg@optimus.ietf.org>; Tue, 30 Apr 2002 11:24:53 -0400 (EDT)
Received: from sm11.texas.rr.com (sm11.texas.rr.com [24.93.35.42])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA24918
	for <tsvwg@ietf.org>; Tue, 30 Apr 2002 11:24:47 -0400 (EDT)
Received: from austin.ibm.com (cs242715-159.austin.rr.com [24.27.15.159])
	by sm11.texas.rr.com (8.12.1/8.12.0) with ESMTP id g3UFIotk020612;
	Tue, 30 Apr 2002 10:18:53 -0500
Message-ID: <3CCEB816.9000506@austin.ibm.com>
Date: Tue, 30 Apr 2002 10:28:22 -0500
From: Jon Grimm <jgrimm@austin.ibm.com>
Organization: IBM Corp.
User-Agent: Mozilla/5.0 (Windows; U; Win98; en-US; rv:0.9.4) Gecko/20011128 Netscape6/6.2.1
X-Accept-Language: en-us
MIME-Version: 1.0
To: Randall Stewart <randall@stewart.chicago.il.us>
CC: Qiaobing Xie <qxie1@email.mot.com>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, bidulock@openss7.org,
        Michael.Tuexen@icn.siemens.de, s_balani@trillium.com,
        mo_khan@trillium.com, h_pillai@trillium.com, jheitz@lucent.com,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFB4@esebe017.NOE.Nokia.com> <3CCD844A.DB1BC960@austin.ibm.com> <3CCD9267.6456A90E@stewart.chicago.il.us> <3CCDA2AD.CB11491@austin.ibm.com> <3CCDB210.E8FA4663@stewart.chicago.il.us> <3CCDD167.8A0DBC43@email.mot.com> <3CCDDF18.9DDD2030@stewart.chicago.il.us> <3CCDED1D.EFC1E6F0@email.mot.com> <3CCDF641.56A4C9AD@stewart.chicago.il.us> <3CCE9A9F.5010703@austin.ibm.com> <3CCEA184.1D1E29FD@stewart.chicago.il.us>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Thanks Randy,
	I think I see where you are coming from through the examples.

Randall Stewart wrote:

> Jon Grimm wrote:
> 
>>Randall Stewart wrote:
>>
>>
>>>Qiaobing Xie wrote:


[snipping a bunch]


>>>>
>>>>
>>>>
>>>No it is managing two different assocaitions.. not endpoints...
>>>
>>>You are always allowed to have an ENDPOINT with a set of
>>>addresses... Associations have to be able to use subsets
>>>of the addresses... period.. otherwise forget IPv6..
>>>
>>>
>>>
>>I can't see if it matters or not, but how about the delete-ip case?  If
>>an endpoint has multiple associations and a subset of the associations
>>support delete-ip.   What should happens to the endpoint?  Is the
>>address really removed from the endpoint?  There are associations still
>>using that address removed from some of the addresses.  I can see where
>>I might want to keep the address in the endpoint until all users of the
>>address have either ACKed the delete or have terminated.  In this case,
>>I wouldn't want to allow a new endpoint to form with that deleted
>>address.  Is this warped thinking?
>>
>>
> Jon:
> 
> well let see... I think the way we handled this is we DID
> remove it from the "endpoint" or listener... that way you
> will NOT accept new INIT's to this.. of course if the 
> address is truely removed from the system its a MOOT point :>
> 
> Now each association has its own list of addresses that are
> in this "present and not deleted state" or "added but not completed"
> state.
> That way if somehow a packet came in to that address we would
> locate the right association.
> 



I think this is what I was looking for.. it looks like it is possible 
for the non-delete-ip association to continue using the address 
(assuming it is not removed from the system).   In this case, the 
association would be using a superset of its endpoints current address 
set.



> I think the big thing here that I have not stated is I don't
> believe we allow you to add/delete an IP address to a association.
> 


> You do a sctp_bindx() against an endpoint yes.
> 
> You have flags that say (default is on in most cases) DO ASCONF
> if you can.. So you get this behavior on all your fanned out
> associations.. but if I am adding an address and some association
> under my endpoint can't.. no big deal it just goes on a negative
> list saying.. don't use this address. (And we are in the kernel
> so we DO control the IP address that goes on the wire... if you
> can't control this address you are probably in rough shape and
> unable to do add-ip .. the more I think about it... of course
> if you can't control your IP address your stack is kinda broken :>)
> 
> But you CAN'T say ... add/delete this address to association X under
> this endpoint.
> 
> Now for the "connected/peeled-off" socket you can do this .. since doing 
> a sctp_bindx() against the connected socket WOULD add it to the
> association.. but this of course does NOT change the listening
> endpoint... so it has no ramifications against future INIT's that
> arrive at the other endpoint that did NOT do a sctp_bindx() on the
> address.
> 


Yeah, but this is where I get a little picky.  An association is a 
function of two endpoints.. a local and a remote.   What is this peeled 
off association's endpoint?  I'm guessing this is why you've migrated to 
talk about listener, as opposed to endpoint??  ;-)   To say an 
association doesn't have an endpoint seems a little odd.  If your 
perspective says that the peeled-off association belongs to a different 
endpoint, then this really becomes just an API implementation issue, 
since any association under an endpoint could migrate to another socket. 
  The association is only loosely tied to the listener endpoint by the 
socket.

Whew, enough of this for me.   I really just wanted to figure out what 
was allowed and what wasn't.   I think I'm there! (assuming I 
interpretted your comments above correctly).   I'm guessing folks will 
still use the endpoint term loosely, but at least we can nail down 
protocol (and possibly Sockets) behavior.

Thanks again,

jon


> 
> R
> 
> 
> 
>><snipped everything else>
>>
> 




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 30 16:14:21 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA07334
	for <tsvwg-archive@odin.ietf.org>; Tue, 30 Apr 2002 16:14:20 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA22705
	for tsvwg-archive@odin.ietf.org; Tue, 30 Apr 2002 16:14:24 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA21056;
	Tue, 30 Apr 2002 15:48:20 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA07642
	for <tsvwg@optimus.ietf.org>; Tue, 30 Apr 2002 12:39:17 -0400 (EDT)
Received: from hermes.fm.intel.com (fmr01.intel.com [192.55.52.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA28200
	for <tsvwg@ietf.org>; Tue, 30 Apr 2002 12:39:13 -0400 (EDT)
Received: from talaria.fm.intel.com (talaria.fm.intel.com [10.1.192.39])
	by hermes.fm.intel.com (8.11.6/8.11.6/d: outer.mc,v 1.42 2002/04/26 23:25:13 root Exp $) with ESMTP id g3UGeB405095
	for <tsvwg@ietf.org>; Tue, 30 Apr 2002 16:40:11 GMT
Received: from fmsmsxvs043.fm.intel.com (fmsmsxv043-1.fm.intel.com [132.233.48.128])
	by talaria.fm.intel.com (8.11.6/8.11.6/d: inner.mc,v 1.17 2002/04/27 00:24:04 root Exp $) with SMTP id g3UGfak13822
	for <tsvwg@ietf.org>; Tue, 30 Apr 2002 16:41:36 GMT
Received: from fmsmsx28.fm.intel.com ([132.233.42.28])
 by fmsmsxvs043.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002043009381801098
 ; Tue, 30 Apr 2002 09:38:18 -0700
Received: by fmsmsx28.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <J9TT599Y>; Tue, 30 Apr 2002 09:39:14 -0700
Message-ID: <F1CE15E08172D4119247009027AE9D50191173B0@FMSMSX37>
From: "Balani, Sandeep" <s_balani@trillium.com>
To: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        Jon Grimm <jgrimm@austin.ibm.com>
Cc: Qiaobing Xie <qxie1@email.mot.com>,
        Qiaobing Xie<Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, bidulock@openss7.org,
        Michael.Tuexen@icn.siemens.de,
        "Balani, Sandeep"<s_balani@trillium.com>,
        "Khan, Mohammad" <mo_khan@trillium.com>,
        "Pillai, Hari" <h_pillai@trillium.com>, jheitz@lucent.com,
        tsvwg@ietf.org
Subject: RE: [Tsvwg] EndPoint and Association Address conflict
Date: Tue, 30 Apr 2002 09:39:05 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

hi all, 
  my comments .. 



> -----Original Message-----
> From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: Tuesday, April 30, 2002 6:52 AM
> To: Jon Grimm
> Cc: Qiaobing Xie; Qiaobing Xie; Ivan.Arias-Rodriguez@nokia.com;
> bidulock@openss7.org; Michael.Tuexen@icn.siemens.de;
> s_balani@trillium.com; mo_khan@trillium.com; h_pillai@trillium.com;
> jheitz@lucent.com; tsvwg@ietf.org
> Subject: Re: [Tsvwg] EndPoint and Association Address conflict
> 
> 
> Jon Grimm wrote:
> > 
> > Randall Stewart wrote:
> > 
> > > Qiaobing Xie wrote:
> > >
> > >>R,
> > >>
> > >>A couple of comments...
> > >>
> > >>Randall Stewart wrote:
> > >>...
> > >>
> > >>>Now I have done a bit more thinking on this. The corner case that
> > >>>we were worried about was this (if I remember right):
> > >>>
> > >>>Host A                                          Host Z
> > >>>IP-A/IP-B                                       IP-Z
> > >>>(An UDP model Endpoint
> > >>> Listening on Port 2222)
> > >>>        IP-A:Port-2222<-----------------INIT(From IP-Z:Port 3333)
> > >>>        ----------INIT-ACK(IP-A,IP-B:Port 2222)-------->
> > >>>        <--------------COOKIE-ECHO----------------------
> > >>>        ----------------COOKIE-ACK---------------------->
> > >>>
> > >>>            Now at this point A-1 is up.
> > >>>
> > >>>Now along the line we do
> > >>>
> > >>>ifconfg New-Intf IP-C
> > >>>
> > >>>and magically do
> > >>>
> > >>>--------------ASCONF(ADD-IP:IP-C)------------------------>
> > >>>    <---------------------OperationalError(Unknown Chunk 
> ASCONF)-----
> > >>>
> > >>>So here we have a potential problem...
> > >>>
> > >>At this point, would our local stack (I am avoiding to 
> use endpoint
> > >>here) actually start to use IP-C as the source addr in 
> any outbound
> > >>packets in the current association? If so, that would 
> trigger an abort
> > >>from the peer and kill the current association. If not, 
> then strictly
> > >>speaking, IP-C has not yet been added to the local 
> **endpoint** that the
> > >>current associaiton is using. It has only been made 
> available to the
> > >>local stack for future use.
> > >>
> > >>
> > >>>                    now our silly peer does
> > >>>
> > >>> <------------------------IP-C:2222 INIT(IP-Z:Port 3333)
> > >>>
> > >>>What do we do?
> > >>>
> > >>> --------INIT-ACK(IP-A,IP-B,IP-C: Port 2222)---------->IP-Z:Port
> > >>>3333--->
> > >>>
> > >>When the local stack does this, it starts to manage two logically
> > >>different local endpoints - the one used by the old 
> association that has
> > >>only IP-A and IP-B; the other is intended for the new 
> association, as
> > >>advertised in this INIT-ACK, that has IP-A, IP-B, and IP-C.
> > >>
> > >>
> > >
> > > No it is managing two different assocaitions.. not endpoints...
> > >
> > > You are always allowed to have an ENDPOINT with a set of
> > > addresses... Associations have to be able to use subsets
> > > of the addresses... period.. otherwise forget IPv6..
> > >
> > >
> > 
> > I can't see if it matters or not, but how about the 
> delete-ip case?  If
> > an endpoint has multiple associations and a subset of the 
> associations
> > support delete-ip.   What should happens to the endpoint?  Is the
> > address really removed from the endpoint?  There are 
> associations still
> > using that address removed from some of the addresses.  I 
> can see where
> > I might want to keep the address in the endpoint until all 
> users of the
> > address have either ACKed the delete or have terminated.  
> In this case,
> > I wouldn't want to allow a new endpoint to form with that deleted
> > address.  Is this warped thinking?
> > 
> Jon:
> 
> well let see... I think the way we handled this is we DID
> remove it from the "endpoint" or listener... that way you
> will NOT accept new INIT's to this.. of course if the 
> address is truely removed from the system its a MOOT point :>
> 
> Now each association has its own list of addresses that are
> in this "present and not deleted state" or "added but not completed"
> state.
> That way if somehow a packet came in to that address we would
> locate the right association.
> 
> I think the big thing here that I have not stated is I don't
> believe we allow you to add/delete an IP address to a association.
> 
> You do a sctp_bindx() against an endpoint yes.
> 
> You have flags that say (default is on in most cases) DO ASCONF
> if you can.. So you get this behavior on all your fanned out
> associations.. but if I am adding an address and some association
> under my endpoint can't.. no big deal it just goes on a negative
> list saying.. don't use this address. (And we are in the kernel
> so we DO control the IP address that goes on the wire... if you
> can't control this address you are probably in rough shape and
> unable to do add-ip .. the more I think about it... of course
> if you can't control your IP address your stack is kinda broken :>)
> 
> But you CAN'T say ... add/delete this address to association X under
> this endpoint.
> 
> Now for the "connected/peeled-off" socket you can do this .. 
> since doing 
> a sctp_bindx() against the connected socket WOULD add it to the
> association.. but this of course does NOT change the listening
> endpoint... so it has no ramifications against future INIT's that
> arrive at the other endpoint that did NOT do a sctp_bindx() on the
> address.
> 

[ Sandeep] I think the case u just describe above is perfect case that an
Addition/deletion of IP address is based on an *ENDPOINT* not an association
[ i think that we all agree to that ], although i know that final
beneficiary is Association. 
But we don't need to confuse the readers that ADD/DELETE IP are happening on
an association. 

" We can always say that Addition/Deletion of  IP addresses are allowed on
an Endpoint. And because Source IP Address list is changing all the
association on that Endpoint need to modify their source IP address list and
indicate the same to the peer." 
  
any thoughts....

-thanks
sandeep 
> 
> R
> 
> 
> > 
> > <snipped everything else>
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 30 16:14:38 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA07356
	for <tsvwg-archive@odin.ietf.org>; Tue, 30 Apr 2002 16:14:38 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA22719
	for tsvwg-archive@odin.ietf.org; Tue, 30 Apr 2002 16:14:41 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA20777;
	Tue, 30 Apr 2002 15:46:56 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA15119
	for <tsvwg@optimus.ietf.org>; Tue, 30 Apr 2002 07:16:26 -0400 (EDT)
Received: from CNRI.Reston.VA.US (localhost [127.0.0.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA13271;
	Tue, 30 Apr 2002 07:16:23 -0400 (EDT)
Message-Id: <200204301116.HAA13271@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: IETF-Announce: ;
Cc: tsvwg@ietf.org
From: Internet-Drafts@ietf.org
Reply-to: Internet-Drafts@ietf.org
Date: Tue, 30 Apr 2002 07:16:23 -0400
Subject: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-tfrc-04.txt,.ps
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the Transport Area Working Group Working Group of the IETF.

	Title		: TCP Friendly Rate Control (TFRC):Protocol 
                          Specification
	Author(s)	: M. Handley, J. Padhye, S. Floyd, J. Widmer
	Filename	: draft-ietf-tsvwg-tfrc-04.txt,.ps
	Pages		: 23
	Date		: 29-Apr-02
	
This document specifies TCP-Friendly Rate Control (TFRC).
TFRC is a congestion control mechanism for unicast flows
operating in a best-effort Internet environment.  It is
reasonably fair when competing for bandwidth with TCP flows,
but has a much lower variation of throughput over time
compared with TCP, making it more suitable for applications
such as telephony or streaming media where a relatively smooth
sending rate is of importance.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-tfrc-04.txt

To remove yourself from the IETF Announcement list, send a message to 
ietf-announce-request with the word unsubscribe in the body of the message.

Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-tsvwg-tfrc-04.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-tsvwg-tfrc-04.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<20020429151500.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-tsvwg-tfrc-04.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-tsvwg-tfrc-04.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<20020429151500.I-D@ietf.org>

--OtherAccess--

--NextPart--




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 30 16:15:58 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA07423
	for <tsvwg-archive@odin.ietf.org>; Tue, 30 Apr 2002 16:15:58 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA22830
	for tsvwg-archive@odin.ietf.org; Tue, 30 Apr 2002 16:16:01 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA21113;
	Tue, 30 Apr 2002 15:48:51 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA18401
	for <tsvwg@optimus.ietf.org>; Tue, 30 Apr 2002 15:07:38 -0400 (EDT)
Received: from motgate.mot.com (motgate.mot.com [129.188.136.100])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA04714
	for <tsvwg@ietf.org>; Tue, 30 Apr 2002 15:07:34 -0400 (EDT)
Received: [from pobox2.mot.com (pobox2.mot.com [136.182.15.8]) by motgate.mot.com (motgate 2.1) with ESMTP id MAA21134; Tue, 30 Apr 2002 12:07:37 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox2.mot.com (MOT-pobox2 2.0) with ESMTP id MAA11474; Tue, 30 Apr 2002 12:07:36 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g3UJ7Xb11660;
	Tue, 30 Apr 2002 14:07:33 -0500 (CDT)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id OAA23044;
	Tue, 30 Apr 2002 14:07:43 -0500 (CDT)
Message-ID: <3CCEEC2D.30B0A7C0@email.mot.com>
Date: Tue, 30 Apr 2002 14:10:37 -0500
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: "Balani, Sandeep" <s_balani@trillium.com>
CC: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        Jon Grimm <jgrimm@austin.ibm.com>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, bidulock@openss7.org,
        Michael.Tuexen@icn.siemens.de, "Khan, Mohammad" <mo_khan@trillium.com>,
        "Pillai, Hari" <h_pillai@trillium.com>, jheitz@lucent.com,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <F1CE15E08172D4119247009027AE9D50191173B0@FMSMSX37>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi, all,

This discussion is going nowhere unless we all can agree on what we are
talking about :-< 

There are at least two parallel lines of thinking (i.e., socket
API-focused and protocol definition-focused) and three different views
of what an "endpoint" is (host-view, socket-view, association-view). We
don't even have an agreement on how these views differ (or whether they
actually differ), let alone what rules should be formulated. 

I hope someone can name these "things" first. Without these terms
clarified to all the participants, the discussion is IMO just a waste of
everybody's time.

-Qiaobing


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 30 16:19:34 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA07626
	for <tsvwg-archive@odin.ietf.org>; Tue, 30 Apr 2002 16:19:33 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA23055
	for tsvwg-archive@odin.ietf.org; Tue, 30 Apr 2002 16:19:37 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA20820;
	Tue, 30 Apr 2002 15:47:22 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA22950
	for <tsvwg@optimus.ietf.org>; Tue, 30 Apr 2002 09:19:11 -0400 (EDT)
Received: from sm13.texas.rr.com (sm13.texas.rr.com [24.93.35.40])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA19096
	for <tsvwg@ietf.org>; Tue, 30 Apr 2002 09:19:07 -0400 (EDT)
Received: from austin.ibm.com (cs242715-159.austin.rr.com [24.27.15.159])
	by sm13.texas.rr.com (8.12.0.Beta16/8.12.0.Beta16) with ESMTP id g3UDRvSI019081;
	Tue, 30 Apr 2002 08:28:00 -0500
Message-ID: <3CCE9A9F.5010703@austin.ibm.com>
Date: Tue, 30 Apr 2002 08:22:39 -0500
From: Jon Grimm <jgrimm@austin.ibm.com>
Organization: IBM Corp.
User-Agent: Mozilla/5.0 (Windows; U; Win98; en-US; rv:0.9.4) Gecko/20011128 Netscape6/6.2.1
X-Accept-Language: en-us
MIME-Version: 1.0
To: Randall Stewart <randall@stewart.chicago.il.us>
CC: Qiaobing Xie <qxie1@email.mot.com>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, bidulock@openss7.org,
        Michael.Tuexen@icn.siemens.de, s_balani@trillium.com,
        mo_khan@trillium.com, h_pillai@trillium.com, jheitz@lucent.com,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] EndPoint and Association Address conflict
References: <58AEDE27D48F884285007ABF827FF26308EFB4@esebe017.NOE.Nokia.com> <3CCD844A.DB1BC960@austin.ibm.com> <3CCD9267.6456A90E@stewart.chicago.il.us> <3CCDA2AD.CB11491@austin.ibm.com> <3CCDB210.E8FA4663@stewart.chicago.il.us> <3CCDD167.8A0DBC43@email.mot.com> <3CCDDF18.9DDD2030@stewart.chicago.il.us> <3CCDED1D.EFC1E6F0@email.mot.com> <3CCDF641.56A4C9AD@stewart.chicago.il.us>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit



Randall Stewart wrote:

> Qiaobing Xie wrote:
> 
>>R,
>>
>>A couple of comments...
>>
>>Randall Stewart wrote:
>>...
>>
>>>Now I have done a bit more thinking on this. The corner case that
>>>we were worried about was this (if I remember right):
>>>
>>>Host A                                          Host Z
>>>IP-A/IP-B                                       IP-Z
>>>(An UDP model Endpoint
>>> Listening on Port 2222)
>>>        IP-A:Port-2222<-----------------INIT(From IP-Z:Port 3333)
>>>        ----------INIT-ACK(IP-A,IP-B:Port 2222)-------->
>>>        <--------------COOKIE-ECHO----------------------
>>>        ----------------COOKIE-ACK---------------------->
>>>
>>>            Now at this point A-1 is up.
>>>
>>>Now along the line we do
>>>
>>>ifconfg New-Intf IP-C
>>>
>>>and magically do
>>>
>>>--------------ASCONF(ADD-IP:IP-C)------------------------>
>>>    <---------------------OperationalError(Unknown Chunk ASCONF)-----
>>>
>>>So here we have a potential problem...
>>>
>>At this point, would our local stack (I am avoiding to use endpoint
>>here) actually start to use IP-C as the source addr in any outbound
>>packets in the current association? If so, that would trigger an abort
>>from the peer and kill the current association. If not, then strictly
>>speaking, IP-C has not yet been added to the local **endpoint** that the
>>current associaiton is using. It has only been made available to the
>>local stack for future use.
>>
>>
>>>                    now our silly peer does
>>>
>>> <------------------------IP-C:2222 INIT(IP-Z:Port 3333)
>>>
>>>What do we do?
>>>
>>> --------INIT-ACK(IP-A,IP-B,IP-C: Port 2222)---------->IP-Z:Port
>>>3333--->
>>>
>>When the local stack does this, it starts to manage two logically
>>different local endpoints - the one used by the old association that has
>>only IP-A and IP-B; the other is intended for the new association, as
>>advertised in this INIT-ACK, that has IP-A, IP-B, and IP-C.
>>
>>
> 
> No it is managing two different assocaitions.. not endpoints...
> 
> You are always allowed to have an ENDPOINT with a set of
> addresses... Associations have to be able to use subsets
> of the addresses... period.. otherwise forget IPv6..
> 
> 


I can't see if it matters or not, but how about the delete-ip case?  If 
an endpoint has multiple associations and a subset of the associations 
support delete-ip.   What should happens to the endpoint?  Is the 
address really removed from the endpoint?  There are associations still 
using that address removed from some of the addresses.  I can see where 
I might want to keep the address in the endpoint until all users of the 
address have either ACKed the delete or have terminated.  In this case, 
I wouldn't want to allow a new endpoint to form with that deleted 
address.  Is this warped thinking?




Best Regards,
Jon

<snipped everything else>



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 30 16:39:21 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA08242
	for <tsvwg-archive@odin.ietf.org>; Tue, 30 Apr 2002 16:39:21 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA24309
	for tsvwg-archive@odin.ietf.org; Tue, 30 Apr 2002 16:39:25 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA23109;
	Tue, 30 Apr 2002 16:20:06 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA23082
	for <tsvwg@optimus.ietf.org>; Tue, 30 Apr 2002 16:20:02 -0400 (EDT)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-230.dallas.net [209.44.42.230])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA07650
	for <tsvwg@ietf.org>; Tue, 30 Apr 2002 16:19:57 -0400 (EDT)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id PAA14492;
	Tue, 30 Apr 2002 15:19:26 -0500
Date: Tue, 30 Apr 2002 15:19:26 -0500
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Randall Stewart <randall@stewart.chicago.il.us>
Cc: Qiaobing Xie <qxie1@email.mot.com>,
        Michael Tuexen <Michael.Tuexen@icn.siemens.de>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, s_balani@trillium.com,
        mo_khan@trillium.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
Message-ID: <20020430151926.B2547@openss7.org>
Reply-To: bidulock@openss7.org
References: <875228B9-5BAC-11D6-ACAA-0030654C1AB6@icn.siemens.de> <3CCDDF38.6EA2990E@email.mot.com> <3CCDE3BC.A623629@stewart.chicago.il.us> <20020429195039.J15434@openss7.org> <3CCDFA75.E6862A7@stewart.chicago.il.us> <20020429234827.C32335@openss7.org> <3CCE88F3.16AAA93E@stewart.chicago.il.us> <20020430131618.A2547@openss7.org> <3CCEE1E9.4361187B@stewart.chicago.il.us>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <3CCEE1E9.4361187B@stewart.chicago.il.us>; from randall@stewart.chicago.il.us on Tue, Apr 30, 2002 at 01:26:50PM -0500
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Randall,

The mistake is returning the error on bind when SO_REUSEADDR is
set.  EADDRINUSE is returned on the connect() instead.  When
ADD-IP goes to add an IP address to an existing association, it
can check for address conflicts in the established hashes and


On Tue, 30 Apr 2002, Randall Stewart wrote:

> "Brian F. G. Bidulock" wrote:
> 
> > These endpoints are not listening.  When they engage in fully
> > formed connections the endpoint is determined from the established
> > hashes which requires a match on vtag, source address, source port,
> > destination address and destination port.  As long as this set
> > uniquely identifies an endpoint (and listen() and connect() will
> > make sure that this is true) there is never any problem, just
> > as there is never any problem for TCP above.
> > 
> 
> Vtags have never been defined as part of the identification
> of an SCTP association. They are used in place of the
> pseudo header and as a defense against blind attacks.. not
> as part of association discrimination... Sorry this won't work.

I use vtag for fast hash lookup to find the entries to make the
daddr/dport/saddr/sport match for established associations for
all chunks not containing a vtag (vtag == 0) or without the T-bit
set.  When the T-bit is set I use the peer tag on a separate peer
tag hash.  For INIT chunks I check the listening hash.

> Two associations are not allowed in this case.. just like
> your TCP description would return a INUSE error...

listen() and connect() ensure that there are no conflicts in the
hashes. (not bind()).  There never results two associations between
the same subset of destination transport addresses per the "MUST NOT"
on Page 13.  It is the statement related to bind (naming an endpoint)
which is incorrect.

> 
> One can easily prevent the TCP case.. i.e. at the connect
> check the hash.... I guess the ADD-IP case also can
> be prevented when one tries to add the IP-C to the second
> association you would simply get the same error :>
> 
> Thinking about it.. we do this now.. since when you attempt
> the sctp_bindx you would get such an error :>

This is the wrong place because of SO_REUSEADDR.  Bind should
permit reuse of local addresses.  It is connect() and listen()
that ensure that no conflicts occur.

--brian

> 
> R
> 
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Apr 30 16:59:39 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA08782
	for <tsvwg-archive@odin.ietf.org>; Tue, 30 Apr 2002 16:59:39 -0400 (EDT)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA25126
	for tsvwg-archive@odin.ietf.org; Tue, 30 Apr 2002 16:59:43 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA24033;
	Tue, 30 Apr 2002 16:32:38 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA24003
	for <tsvwg@optimus.ietf.org>; Tue, 30 Apr 2002 16:32:35 -0400 (EDT)
Received: from stewart.chicago.il.us (user166.64.47.24.dsli.com [64.47.24.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA08035
	for <tsvwg@ietf.org>; Tue, 30 Apr 2002 16:32:30 -0400 (EDT)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewart.chicago.il.us (8.11.1/8.11.1) with ESMTP id g3UKW8H62957;
	Tue, 30 Apr 2002 15:32:08 -0500 (CDT)
	(envelope-from randall@stewart.chicago.il.us)
Message-ID: <3CCEFF48.501E797D@stewart.chicago.il.us>
Date: Tue, 30 Apr 2002 15:32:08 -0500
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: bidulock@openss7.org
CC: Qiaobing Xie <qxie1@email.mot.com>,
        Michael Tuexen <Michael.Tuexen@icn.siemens.de>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        Ivan.Arias-Rodriguez@nokia.com, s_balani@trillium.com,
        mo_khan@trillium.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] End point and association Address conflict
References: <875228B9-5BAC-11D6-ACAA-0030654C1AB6@icn.siemens.de> <3CCDDF38.6EA2990E@email.mot.com> <3CCDE3BC.A623629@stewart.chicago.il.us> <20020429195039.J15434@openss7.org> <3CCDFA75.E6862A7@stewart.chicago.il.us> <20020429234827.C32335@openss7.org> <3CCE88F3.16AAA93E@stewart.chicago.il.us> <20020430131618.A2547@openss7.org> <3CCEE1E9.4361187B@stewart.chicago.il.us> <20020430151926.B2547@openss7.org>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

"Brian F. G. Bidulock" wrote:
> 
> Randall,
> 
> The mistake is returning the error on bind when SO_REUSEADDR is
> set.  EADDRINUSE is returned on the connect() instead.  When
> ADD-IP goes to add an IP address to an existing association, it
> can check for address conflicts in the established hashes and
> 
> On Tue, 30 Apr 2002, Randall Stewart wrote:
> 
> > "Brian F. G. Bidulock" wrote:
> >
> > > These endpoints are not listening.  When they engage in fully
> > > formed connections the endpoint is determined from the established
> > > hashes which requires a match on vtag, source address, source port,
> > > destination address and destination port.  As long as this set
> > > uniquely identifies an endpoint (and listen() and connect() will
> > > make sure that this is true) there is never any problem, just
> > > as there is never any problem for TCP above.
> > >
> >
> > Vtags have never been defined as part of the identification
> > of an SCTP association. They are used in place of the
> > pseudo header and as a defense against blind attacks.. not
> > as part of association discrimination... Sorry this won't work.
> 
> I use vtag for fast hash lookup to find the entries to make the
> daddr/dport/saddr/sport match for established associations for
> all chunks not containing a vtag (vtag == 0) or without the T-bit
> set.  When the T-bit is set I use the peer tag on a separate peer
> tag hash.  For INIT chunks I check the listening hash.
> 

Vtag may be used for fast lookup but it is NOT a association
discriminator. 


> > Two associations are not allowed in this case.. just like
> > your TCP description would return a INUSE error...
> 
> listen() and connect() ensure that there are no conflicts in the
> hashes. (not bind()).  There never results two associations between
> the same subset of destination transport addresses per the "MUST NOT"
> on Page 13.  It is the statement related to bind (naming an endpoint)
> which is incorrect.
> 

Not true when the SO_REUSE flag is NOT set...

> >
> > One can easily prevent the TCP case.. i.e. at the connect
> > check the hash.... I guess the ADD-IP case also can
> > be prevented when one tries to add the IP-C to the second
> > association you would simply get the same error :>
> >
> > Thinking about it.. we do this now.. since when you attempt
> > the sctp_bindx you would get such an error :>
> 
> This is the wrong place because of SO_REUSEADDR.  Bind should
> permit reuse of local addresses.  It is connect() and listen()
> that ensure that no conflicts occur.
> 

Sorry you must use it to filter out such events if you
are going to use it to trigger the ASCONF.... If you want to
NOT use it to trigger the ASCONF and use something else.. then
fine.. you can use that "new-call()" that would generate the
ASCONF to return the error...

R

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



